` |
| OAuth discovery | `https://www.taskade.com/.well-known/oauth-protected-resource/mcp` |
+| API catalog (RFC 9727) | `https://www.taskade.com/.well-known/api-catalog`. It links this server, the REST API v1 spec, and the Action API v2 spec |
## What you can do
@@ -141,6 +142,8 @@ Any other agent can read the Taskade skill at `https://www.taskade.com/SKILL.md`
npx skills add https://www.taskade.com
```
+The skill also tells the agent how to read any Taskade page as Markdown. The agent sends `Accept: text/markdown` to the normal page URL.
+
To add only the server, run one command:
```bash
@@ -310,7 +313,7 @@ Use the Taskade tools to list my workspaces. For each one, show the name and my
You are done when the client calls `list_spaces` and shows your paid workspaces. The list comes in pages, and a page can be empty while more pages remain. If no page lists a workspace, no workspace of this account has MCP access.
{% hint style="info" %}
-**Note:** A `HEAD` request or `curl -I` to `https://www.taskade.com/mcp` returns `401` with `content-type: application/json` and a `WWW-Authenticate: Bearer` header. That response is expected, because the probe sends no token. Do not change the server URL.
+**Note:** A `HEAD` request or `curl -I` to `https://www.taskade.com/mcp` returns `401` with `content-type: application/json` and a `WWW-Authenticate: Bearer` header. This response is normal, because the probe sends no token. Do not change the server URL.
{% endhint %}
| You send | You get |
@@ -327,7 +330,14 @@ Most tools mirror an [Action API v2](api-v2-reference.md) operation. The MCP-nat
Every tool states the MCP hints `readOnlyHint`, `destructiveHint`, and `openWorldHint`. A client can use the hints to approve read tools without a prompt and to ask before a write. Three tools state `openWorldHint: true`, because their effects reach outside the workspace: `prompt_agent`, `run_flow`, and `publish_kit_version`.
-Most tools take a `spaceId`. Call `list_spaces` first to find it. That list is **plan-filtered and paginated**, so a page can be empty while more pages remain. Keep paginating until `hasNextPage` is false.
+The server also adds two Taskade keys to the `_meta` of a tool. A client that does not know them ignores them.
+
+| `_meta` key | Tools | Use it to |
+| --- | --- | --- |
+| `taskade/write: true` | Every tool that is not read-only | Ask the user before a write |
+| `taskade/billed: true` | `prompt_agent` and `run_flow` | Ask the user before a call that uses AI credits or counts toward the monthly automation runs |
+
+Most tools take a `spaceId`. Call `list_spaces` first to find it. That list is **plan-filtered and paginated**, so a page can be empty when more pages remain. Request the next page until `hasNextPage` is false.
### Browse & app editing
@@ -374,7 +384,7 @@ These tools mirror the [Action API v2](api-v2-reference.md) task operations:
| `set_task_field_value` / `delete_task_field_value` | Set or clear a custom field value on a task |
| `delete_task` | Delete a task |
-If Google Calendar is connected on an eligible workspace, dates set with `set_task_date` can sync the same way as dates set in Taskade.
+If an eligible workspace has a Google Calendar connection, a date from `set_task_date` can sync the same way as a date that you set in Taskade.
### Create & orchestrate
@@ -403,7 +413,7 @@ You can change this setting later with `update_agent`. This setting also enables
`run_flow` starts a real run. The run counts toward the monthly automation runs of the plan, and AI steps use credits. A step can also send email or act in a connected app. Call it once for each run that you want.
{% endhint %}
-`run_flow` returns before the run ends. Call `get_flow_runs` to read the result. A run status is `running`, `completed`, `failed`, or `filtered`. A `filtered` run stopped early on purpose and did not fail. If `run_flow` returns the status `unknown` or `running`, do not start the run again. Read it with `get_flow_runs`.
+`run_flow` returns before the run ends. Call `get_flow_runs` to read the result. A run status is `running`, `waiting`, `completed`, `failed`, or `filtered`. A `waiting` run is paused on an approval step until a person decides, or until the step times out. Its `endedAt` is `null`, the same as a `running` run. A `filtered` run stopped early on purpose and did not fail. If `run_flow` returns the status `unknown`, `running`, or `waiting`, do not start the run again. Read it with `get_flow_runs`.
### App kits
@@ -417,7 +427,7 @@ An app kit is a Taskade Genesis app that you published so other people can add a
Only the workspace owner can call `publish_kit_version`. `list_my_kits` lists kits from workspaces that you own, and it hides a kit whose workspace has no MCP access. Archived and removed kits are not listed.
{% hint style="warning" %}
-`publish_kit_version` releases the app as it is now. People who add the kit from then on get this version, and people who already added it keep their copy. From version 2 on, buyers get one email a day at most. Only the newest 20 versions keep a saved copy, so each call removes the saved copy of an older version past that limit.
+`publish_kit_version` releases the app as it is now. People who add the kit from then on get this version, and people who already added it keep their copy. From version 2 on, buyers get one email a day at most. Only the newest 20 versions keep a saved copy. Each call removes the saved copy of an older version past that limit.
{% endhint %}
A paid kit keeps its price. You can publish 5 versions of one kit in an hour.
@@ -505,17 +515,51 @@ Match the message that your client shows to a fix.
| `MCP access is not available on this workspace's current plan.` | The workspace is on the Free plan | Upgrade the workspace to a paid plan, or pick a paid workspace. |
| `Access denied: you are not a member of this space.` | Your account is not a member of the workspace | Make sure that the workspace ID is correct, or ask for an invitation. |
| `Access denied: your role "…" does not have write access. Requires collaborator or above.` | `write_file` needs the Editor role or above | Ask the workspace owner for the Editor role. A read-only role can still browse. |
-| `list_spaces` returns an empty page | The list comes in pages, and it hides workspaces without MCP access | Keep paging until `hasNextPage` is `false`. |
+| `list_spaces` returns an empty page | The list comes in pages, and it hides workspaces without MCP access | Request the next page until `hasNextPage` is `false`. |
| `You do not have permission to run automations in this workspace.` | Your role cannot run automations in this workspace | Ask the workspace owner for a role that can run automations. |
| `"…" is Off. Turn it On in Taskade, then run it again.` | `run_flow` starts only an automation that is On | Turn the automation On in Taskade, then call `run_flow` again. |
| `run_flow starts only an automation that starts "When asked" (a manual trigger).` | The automation starts on another trigger | Change the trigger to **Agent Tool** (the "When asked" trigger), or let the trigger start the automation. |
| `Too many agent prompts in a short time, so the agent did not run. Try again in about … min.` | The account sent 60 `prompt_agent` calls in 10 minutes | Wait the number of minutes in the message, then call `prompt_agent` again. Do not call it in a loop. |
+| `Failed to prompt agent`, with the error code `credits_exhausted` | The workspace has no AI credits left | Upgrade the plan in Taskade, or add credits on a paid plan. A retry does not help. |
+| `Too many runs of "…" in a short time, so nothing ran. Try again in about … min.` | The account started this automation 30 times in 10 minutes | Wait the number of minutes in the message, then call `run_flow` once. Do not call it in a loop. |
+| `Nothing ran. This workspace used all of its … for this month on the … plan.` | `run_flow` reached the monthly limit of automation runs or steps for the plan | Wait for the limit to reset, or upgrade the plan. |
+| `…, so "…" did not run.`, with the error code `connector_plan_required` | A step of the automation uses a connector that the plan does not include | Upgrade the workspace plan, or remove that step. |
| `Only the workspace owner can publish a kit version.` | `publish_kit_version` needs the Owner role in the workspace of the kit | Ask the workspace owner to publish, or to run the tool from their account. |
| `No live kit listing with this id. An archived or removed listing takes no new version.` | The `publicSpaceAppId` is wrong, or the kit is archived or removed | Call `list_my_kits` and use a `publicSpaceAppId` from its result. |
| `Too many versions of this kit in a short time, so nothing was published. Try again in about … min.` | You published 5 versions of this kit in the last hour | Wait the number of minutes in the message, then publish once. |
| The OAuth sign-in starts again and again | The browser is signed in to another Taskade account, or the stored credential is old | Sign in to the right Taskade account in your browser. Remove the Taskade server from the client, then add it again. |
| `401` from a header-based client | The personal access token is wrong or revoked | Create a new token, and update the `Authorization` header. |
+### Read a tool error in code
+
+A tool error is a result with `isError: true`. The text block holds the message from the table above. Many refusals also carry `structuredContent.error`, so a script or an agent can branch without a text match:
+
+```json
+{
+ "isError": true,
+ "structuredContent": {
+ "error": {
+ "class": "rate",
+ "code": "rate_limited",
+ "message": "Too many agent prompts in a short time, so the agent did not run. Try again in about 3 min.",
+ "retry": "after_wait",
+ "retryAfterSec": 142
+ }
+ }
+}
+```
+
+| Field | Value |
+| --- | --- |
+| `class` | `auth` (membership or role), `plan` (the plan has no access), `input` (fix the call), `refused` (the target is not in a state to run), `rate` (too many calls), `limit` (a usage or size limit), or `error` |
+| `code` | A short code, for example `plan_wall`, `not_member`, `forbidden`, `flow_off`, `not_manual`, `connector_plan_required`, `over_limit`, `rate_limited`, `credits_exhausted`, or `too_large` |
+| `message` | The same text as the text block |
+| `next` | Optional. Text to show the user, for example the upgrade step |
+| `retry` | `no` (the same call fails again), `safe` (you can call again now), or `after_wait` (call again after `retryAfterSec` seconds) |
+| `retryAfterSec` | Optional. The number of seconds to wait before you call again. An `after_wait` error carries it |
+
+An error with no `structuredContent.error` is a general failure. Read its text.
+
## Set this up with your coding agent
Paste this block into your coding agent to connect it.
@@ -528,7 +572,7 @@ Connect to the hosted Taskade MCP server at https://www.taskade.com/mcp over Str
- [Which Taskade MCP do I want?](mcp-overview.md): compare hosted MCP, local Workspace MCP, and the MCP Client connector.
- [Workspace MCP](workspace-mcp.md): run a local server that wraps REST API v1.
-- [Run your live app with AI](run-your-app-with-ai.md): run a Taskade Genesis app by talking to Claude.
+- [Run your live app with AI](run-your-app-with-ai.md): run a Taskade Genesis app from a chat with Claude.
- [Action API guide](api-v2-reference.md): the API that most hosted MCP tools mirror.
- [Authentication](developers/authentication.md): personal tokens and OAuth 2.0 in detail.
diff --git a/apis-living-system-development/workspace-mcp-advanced.md b/apis-living-system-development/workspace-mcp-advanced.md
index a406f2b..b47e976 100644
--- a/apis-living-system-development/workspace-mcp-advanced.md
+++ b/apis-living-system-development/workspace-mcp-advanced.md
@@ -157,7 +157,7 @@ When the remote server requires a token, add the optional bearer token.
**An agent cannot reach an MCP server on its own.** Taskade has no per-agent MCP connector settings on any plan.
-To let an agent use a remote MCP tool, configure the agent to trigger an automation that carries the MCP Client step. The automation hands the result back.
+To let an agent use a remote MCP tool, use an automation. Configure the agent to trigger an automation that has the MCP Client step. The automation gives the result back.
The connector applies these connection rules:
@@ -165,7 +165,7 @@ The connector applies these connection rules:
- It rejects credentials embedded in the URL.
- It rejects `localhost`, `*.localhost`, `*.local`, and `*.internal`.
- It rejects loopback, private, link-local, CGNAT, unspecified, and IPv4-mapped internal IP literals.
-- It rejects redirects instead of following them.
+- It rejects redirects and does not follow them.
- It gives each transport 20 seconds to complete the handshake.
See **[MCP Client connector](https://www.taskade.com/learn/connect/mcp-client-integration)** for setup. See **[Which Taskade MCP do I want?](mcp-overview.md)** to compare directions.
@@ -186,19 +186,19 @@ For services without MCP, use the [100+ native integrations](https://www.taskade
| Tool timeout | Large response or slow upstream | Examine the upstream service. Reduce the query scope |
-If the problem stays, file an issue at [github.com/taskade/taskade/issues](https://github.com/taskade/taskade/issues) with your MCP logs.
+If the problem stays, file an issue at [github.com/taskade/mcp/issues](https://github.com/taskade/mcp/issues) with your MCP logs. That repository holds the source of `@taskade/mcp-server`.
## Security Best Practices
- **Audit tool exposure for public agents.** Opt sensitive tools (file access, automation triggers) out of public agent configurations.
-- **Assume a leaked token is a full account compromise.** Personal access tokens carry no workspace or resource scope. A leak exposes every workspace the owning account can reach. Revoke the token at [settings/api](https://www.taskade.com/settings/api) instead of reasoning about the blast radius.
+- **Assume a leaked token is a full account compromise.** Personal access tokens carry no workspace or resource scope. A leak exposes every workspace that the account of the token can reach. Revoke the token at [settings/api](https://www.taskade.com/settings/api). Do not try to measure the damage first.
- **Isolate by account, not by token.** If an integration must access one workspace, give it a dedicated Taskade account.
- Invite that account only to the required workspace. Then issue the token from that account.
- **Rotate on every personnel change.** When a teammate leaves, rotate any shared tokens.
- **Monitor the workspace activity log** for unexpected MCP-initiated actions.
-- The hosted Taskade MCP is always served over TLS at `https://www.taskade.com/mcp`.
+- Taskade always serves the hosted Taskade MCP over TLS at `https://www.taskade.com/mcp`.
## Set this up with your coding agent
@@ -213,7 +213,7 @@ Start the local Taskade MCP server with "npx -y @taskade/mcp-server" over stdio.
## Next steps
-- [Workspace MCP](workspace-mcp.md): install the local server and set up your client.
+- [Workspace MCP](workspace-mcp.md): install the local server and configure your client.
- [Hosted Taskade MCP](genesis-app-mcp.md): the no-install server with OAuth.
- [Action API guide](api-v2-reference.md): the API behind the v2 tools.
- [MCP Client connector](https://www.taskade.com/learn/connect/mcp-client-integration): call another MCP server from an automation.
diff --git a/genesis-living-system-builder/ai-features/README.md b/genesis-living-system-builder/ai-features/README.md
index 1cbd0fa..d490d5a 100644
--- a/genesis-living-system-builder/ai-features/README.md
+++ b/genesis-living-system-builder/ai-features/README.md
@@ -5,7 +5,7 @@ description: >-
# AI Features Overview
-AI features let you build agents that reason over your workspace, learn from your knowledge, use tools, and work together in teams.
+With AI features, you can make AI agents that read your workspace, answer from your knowledge, use tools, and work together in teams.
## I want to…
@@ -17,36 +17,36 @@ AI features let you build agents that reason over your workspace, learn from you
| Build a team of agents | [AI Agent Teams](agent-teams.md) |
| Run multiple agents together | [Multi-Agent Collaboration](multi-agents.md) |
-**Think. Learn. Adapt.** Taskade AI features form the living intelligence layer of Workspace DNA, creating systems that understand, evolve, and provide increasingly intelligent value with every interaction.
+AI features are the Intelligence layer of Workspace DNA. Your agents use your projects and files to answer questions and do work.
{% hint style="success" %}
-**Living Intelligence Layer** — AI that doesn't just assist, but learns from your business, adapts to your needs, and continuously improves. Powered by Taskade EVE, your living AI companion.
+**Intelligence layer:** Your agents answer from your business knowledge and obey your instructions. Taskade EVE, the AI assistant of Taskade, helps you build and use them.
{% endhint %}
-## What is Living Intelligence?
+## What is the Intelligence layer?
-Living Intelligence is the agent layer in Taskade. It’s how your workspace can **reason**, **draft**, **classify**, and **take action** using context from your projects.
+The Intelligence layer is the agent layer in Taskade. With it, your workspace can **reason**, **draft**, **classify**, and **take action** with the context of your projects.
-Canonical foundation: [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
+Read the foundation first: [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
-### What you use it for
+### What you can do with it
-* **Customer-facing help**: answer questions, route requests, propose next steps.
-* **Ops + admin**: summarize updates, draft emails, generate reports, fill data gaps.
-* **Sales + marketing**: qualify leads, personalize outreach, reuse brand voice.
-* **Research**: extract key points from docs, compare options, track decisions.
+* **Help customers:** Answer questions, send requests to the right person, and suggest next steps.
+* **Do operations work:** Summarize updates, draft emails, make reports, and fill gaps in data.
+* **Support sales and marketing:** Qualify leads, personalize outreach, and write in your brand voice.
+* **Do research:** Find the key points in documents, compare options, and record decisions.
### Core building blocks
#### AI agents
-Agents are persistent assistants. You can scope them to a workspace, project, or use case. Learn more about [Taskade AI Agents](https://www.taskade.com/agents).
+An agent is an AI assistant that stays in your workspace. You can give it one job, one project, or one use case. Read more about [Taskade AI Agents](https://www.taskade.com/agents).
Start here: [AI Agents Getting Started](ai-agents-getting-started.md).
-#### Knowledge (what agents learn from)
+#### Knowledge (what agents answer from)
-Most “smartness” comes from the knowledge you give agents. That includes project data and uploaded docs.
+The quality of the answers of an agent comes mostly from the knowledge that you give it. Knowledge includes project data and uploaded documents. This knowledge is the persistent memory of the agent.
Start here:
@@ -55,37 +55,38 @@ Start here:
#### Tools (what agents can do)
-Tools let agents do more than write text. They can create tasks, call integrations, and move work forward.
+With tools, an agent can do more than write text. It can create tasks, call integrations, and move work forward.
Start here:
* [Integration Overview](../automation/comprehensive-integration-guide.md)
* [Actions & Triggers](../automation/actions.md)
-### Getting started (fast)
+### Get started fast
1. Create one agent for one job.
-2. Give it 2–3 knowledge sources.
+2. Give it 2–3 knowledge sources. On Free, each agent holds one knowledge source. Paid plans hold up to 50.
3. Run it on real inputs.
-4. Tighten prompts and add tools only if needed.
+4. Make the instructions more exact.
+5. If the agent must act, add tools.
-### Where it shows up
+### Where agents work
-* In your workspace as an assistant.
-* Inside Taskade Genesis apps as an AI layer.
-* In automations as a decision-maker.
+* In your workspace, as an assistant
+* Inside Taskade Genesis apps, as the AI layer
+* In automations, as the step that makes decisions
## Next steps
-* [AI Agents Getting Started](ai-agents-getting-started.md) — create one agent for one job
-* [Tools for AI Agents](agent-tools.md) — let agents take action, not just write text
-* [Multi-Agent Collaboration](multi-agents.md) — coordinate agents on bigger work
-* [Taskade Genesis Overview](../genesis/) — build apps with agents
-* [Automations Overview](../automation/) — put actions on autopilot
-* [taskade.com/agents](https://www.taskade.com/agents) — browse agent templates
+* [AI Agents Getting Started](ai-agents-getting-started.md): create one agent for one job.
+* [Tools for AI Agents](agent-tools.md): let agents take action, not only write text.
+* [Multi-Agent Collaboration](multi-agents.md): let agents work together on larger work.
+* [Taskade Genesis Overview](../genesis/): build apps with agents.
+* [Automations Overview](../automation/): let actions run on their own.
+* [taskade.com/agents](https://www.taskade.com/agents): browse agent templates.
{% hint style="info" %}
-This page stays overview-level. Model comparisons, tool catalogs, and deep agent training patterns live in the linked guides.
+This page is an overview. The linked guides give model comparisons, tool lists, and detailed patterns for agent knowledge.
{% endhint %}
---
diff --git a/genesis-living-system-builder/ai-features/agent-knowledge.md b/genesis-living-system-builder/ai-features/agent-knowledge.md
index 77fd2a0..7d83558 100644
--- a/genesis-living-system-builder/ai-features/agent-knowledge.md
+++ b/genesis-living-system-builder/ai-features/agent-knowledge.md
@@ -9,23 +9,26 @@ description: >-
## Overview
-Agent knowledge is the information your custom AI agents use to provide **contextual, accurate responses**. It combines pre-trained AI capabilities with **your own data** — documents, URLs, projects, and media — to create agents that truly understand your business.
+Agent knowledge is the information that your custom AI agents answer from. It is **your own data**: documents, links, projects, and media. With this data, the agent gives **accurate answers about your business** instead of general answers. Knowledge is the persistent memory of the agent.
-> **Workspace DNA connection:** Knowledge is what transforms a generic AI into a specialist. The more knowledge you add, the richer your [Workspace DNA](../genesis/how-genesis-works.md) becomes, and the smarter all your Taskade Genesis apps get.
+> **Workspace DNA connection:** Knowledge makes a general AI into a specialist for your work. Your knowledge is the Memory part of [Workspace DNA](../genesis/how-genesis-works.md). Your Taskade Genesis apps use the same Memory.
---
## Explore the Knowledge Tab
-1. Go to the **Agents** tab.
-2. Click the three dots (⋯) next to your agent.
-3. Select **Edit agent**.
-4. Open the **Knowledge** tab.
-5. Turn on the **Enable knowledge** switch. It is off by default, and you cannot add sources until it is on.
+1. In your workspace sidebar, open **AI Agents**.
+2. Click the agent. You can also click the three dots (⋯) next to it and select **Add knowledge**.
+3. Click the **Knowledge** tab, then the **Sources** sub-tab.
+4. Turn on the **Enable knowledge** switch. The switch is off by default. You cannot add sources until it is on.
-The Knowledge tab shows all sources your agent has been trained on, with options to add, remove, and manage sources.
+The side panel groups the sources: **Taskade Workspace**, **Web Sources**, **Cloud Import**, **Local Upload**, and **Custom Text**. Your sources show under **Trained Sources**. The agent reads a source after its status shows that it is trained.
+
+{% hint style="info" %}
+A Free workspace can add one knowledge source to each agent. Every paid plan holds up to 50 sources for each agent. Web sources and cloud imports need a paid plan.
+{% endhint %}
---
@@ -33,15 +36,15 @@ The Knowledge tab shows all sources your agent has been trained on, with options
### Supported Knowledge Sources
-| Source Type | Formats / Types | How It's Used |
+| Source Type | Formats / Types | What Taskade does with it |
|---|---|---|
-| **Documents** | PDF, CSV, TXT, DOCX, MD, PPTX, XLSX, EPUB | Full text extraction and indexing |
-| **Web URLs** | Any web page, blog post, documentation page | Content scraped and indexed |
-| **YouTube videos** | Video URLs | Transcript extracted and indexed |
-| **Social content** | Tweets, Reddit posts, news articles | Content extracted and indexed |
-| **Cloud files** | Google Drive, Dropbox, Box, OneDrive | File content pulled and indexed |
-| **Taskade projects** | Any project in your workspace | Live, dynamic knowledge (updates automatically) |
-| **Manual sources** | Markdown or plain text created in the Knowledge tab | Direct knowledge injection |
+| **Documents** | PDF, CSV, TXT, DOCX, MD, PPTX, XLSX, XLS, EPUB | Reads all the text and indexes it |
+| **Web URLs** (paid plans) | Website URL, blog or article | Reads the page and indexes it |
+| **Feeds and sitemaps** (paid plans) | RSS or Atom feed, sitemap XML | Reads the content and indexes it |
+| **YouTube videos** (paid plans) | Video URLs | Reads the transcript and indexes it |
+| **Cloud files** (paid plans) | Google Drive, Dropbox, Box, OneDrive | Reads the file content and indexes it |
+| **Taskade projects** | Projects in your workspace | Live knowledge that updates when the project changes |
+| **Manual sources** | Markdown or plain text that you write in the Knowledge tab | Adds your text to the knowledge as it is |
---
@@ -49,83 +52,100 @@ The Knowledge tab shows all sources your agent has been trained on, with options
### Upload Documents
-1. Open the Knowledge tab and make sure that **Enable knowledge** is on.
-2. Click the **ADD SOURCES** box to choose files from your device, or drag & drop them in.
-3. Or click **Add Media** to pick a file already uploaded to your projects.
-4. Click **Update** to save. Files are processed and indexed automatically.
+1. Open the Knowledge tab.
+2. Make sure that **Enable knowledge** is on.
+3. Under **Local Upload**, click **Upload from Computer**. You can also drag the files into the upload box.
+4. To use a file that is already in your projects, click **Add Media** under **Taskade Workspace**.
+
+Each file then shows under **Trained Sources**. Taskade processes and indexes each file.
+
+The Knowledge tab does not accept image files. It does not read text from scanned pages. To teach an agent from a scanned file, convert the file to text first. If you drop a file that Taskade cannot read, Taskade skips it, and a notice names the file. The other files of the batch upload as usual.
**Supported document formats:**
| Format | Best For |
|---|---|
-| **PDF** | Brand guidelines, reports, manuals, legal docs |
-| **CSV / XLSX** | Customer data, product catalogs, pricing tables |
+| **PDF** | Brand guidelines, reports, manuals, legal documents |
+| **CSV / XLSX / XLS** | Customer data, product catalogs, price tables |
| **DOCX** | Procedures, policies, templates |
| **TXT / MD** | Plain text knowledge, FAQ answers, notes |
| **PPTX** | Presentations, training materials |
-| **EPUB** | Books, long-form content |
+| **EPUB** | Books, long documents |
### Use Web Resources
-1. In the Knowledge tab, make sure that **Enable knowledge** is on, then click **Add link** (web page), **Add YouTube** (videos), or add from **Google Drive**, **Dropbox**, **Box**, or **OneDrive**.
-2. Paste the link or pick the file.
-3. Click **Update** to save. Content is scraped/transcribed and indexed.
+Web sources and cloud imports need a paid plan. On Free, they show an upgrade badge.
+
+1. Open the Knowledge tab.
+2. Make sure that **Enable knowledge** is on.
+3. For a web source, pick **Website URL**, **Blog / Article**, **YouTube Video**, **RSS / Atom Feed**, or **Sitemap XML** under **Web Sources**.
+4. For a cloud file, pick **Google Drive**, **Dropbox**, **OneDrive**, or **Box** under **Cloud Import**.
+5. Paste the link, or pick the files.
+6. Confirm the selection.
-| Resource Type | What Gets Indexed |
+Taskade reads, transcribes, and indexes the content.
+
+> **Note:** Taskade reads a link or a sitemap as one page. It does not follow the internal links inside the page. Taskade reads a web source one time, when you add it. After the page changes, click **Retrain** to read it again.
+
+| Resource Type | What Taskade indexes |
|---|---|
-| **Web pages** | Full page text content |
-| **YouTube videos** | Full video transcript |
-| **Google Drive files** | Document text content |
-| **Dropbox / Box / OneDrive files** | Document text content |
-| **Blog posts** | Article text and metadata |
-| **Tweets** | Tweet text and thread |
-| **Reddit posts** | Post and top comments |
-| **News articles** | Full article text |
+| **Web pages** | All the text of the page |
+| **YouTube videos** | The full transcript of the video |
+| **Google Drive files** | The text of the document |
+| **Dropbox / Box / OneDrive files** | The text of the document |
+| **Blog posts** | The text and the metadata of the article |
+| **RSS / Atom feeds** | The content of the feed |
+| **Sitemap XML** | The sitemap page, read as one page |
### Create Sources Manually
-1. Open the Knowledge tab and make sure that **Enable knowledge** is on.
-2. Click **Add Markdown/Text**.
-3. Write or paste Markdown/plain text, then click **Add to Knowledge**.
-4. Click **Update** to save — the source is available to your agent on its next reply.
+1. Open the Knowledge tab.
+2. Make sure that **Enable knowledge** is on.
+3. Under **Custom Text**, click **Write Markdown** or **Paste Text**.
+4. Type or paste your text.
+5. Click **Add to Knowledge**.
-> **Tip:** Use manual sources for FAQ answers, custom instructions, edge case handling, or information not available in any file.
+The agent uses the new source in its next reply.
+
+> **Tip:** A manual source is good for FAQ answers, custom instructions, special cases, and facts that are not in a file.
### Dynamic Knowledge (Projects)
-Connect Taskade projects as **living knowledge sources** that update automatically:
+A connected Taskade project is **live knowledge**. When the project changes, the knowledge of the agent changes too.
+
+1. Open the Knowledge tab.
+2. Make sure that **Enable knowledge** is on.
+3. Under **Taskade Workspace**, click **Add Project**.
+4. Select one or more projects.
-1. Open the Knowledge tab and make sure that **Enable knowledge** is on.
-2. Click **Add Project**.
-3. Select one or more projects.
-4. Projects become live knowledge — agent always has the latest data.
+Each connected project carries the **Auto** badge. The agent always reads the latest data of the project.
-> **Dynamic = real-time.** When you update a project, the agent's knowledge updates automatically. No re-upload needed.
+> **Live updates:** When you edit a connected project, the agent gets the change in its next reply. You do not upload it again.
---
## Bulk Training
-Upload multiple sources simultaneously:
+You can add many sources at the same time:
| Method | How |
|---|---|
-| **Multi-file upload** | Select multiple files in the file picker |
-| **Multi-project connect** | Select multiple projects from your workspace |
-| **Multi-URL paste** | Add multiple links at once |
-| **Drag & drop batch** | Drag a folder or multiple files into the Knowledge tab |
+| **Many files** | Select many files in the file picker |
+| **Many projects** | Select many projects from your workspace |
+| **Many links** | Add many links at the same time |
+| **Drag and drop** | Drag many files into the Knowledge tab |
---
## Look Up References
-When your agent responds, it shows which sources it referenced:
+Each reply of the agent shows the sources that it used:
| Feature | Detail |
|---|---|
-| **Source citations** | Chat displays which knowledge sources were used |
-| **Preview capability** | Click a citation to preview the source content |
-| **Transparency** | Users can verify agent responses against source material |
+| **Source citations** | The chat shows which knowledge sources the agent used |
+| **Preview** | Click a citation to see the content of the source |
+| **Transparency** | Users can compare the reply with the source |
@@ -133,21 +153,21 @@ When your agent responds, it shows which sources it referenced:
## Knowledge Backlinks
-Projects used as agent knowledge show **backlinks** indicating which agents reference them:
+A project that an agent uses as knowledge shows a **backlink** to that agent:
| Feature | Detail |
|---|---|
-| **Backlink indicator** | Projects display a badge showing connected agents |
-| **Bi-directional awareness** | See which agents use a project, and which projects an agent uses |
-| **Impact visibility** | Know which agents are affected when you edit a project |
+| **Backlink indicator** | The project shows a badge with the connected agents |
+| **Two-way view** | You can see the agents that use a project, and the projects that an agent uses |
+| **Impact** | You know which agents change when you edit a project |
-> **Tip:** Backlinks help you maintain knowledge integrity. Before editing a project, check its backlinks to understand which agents depend on it.
+> **Tip:** Before you edit a project, look at its backlinks. They show which agents use the project.
---
-## What Projects Expose to Agents
+## What Projects Share with Agents
-When a project is connected as knowledge, agents can access:
+A connected project gives the agent some of its content, but not all of it:
| Accessible | Not Accessible |
|---|---|
@@ -164,26 +184,26 @@ When a project is connected as knowledge, agents can access:
| Practice | Why | Example |
|---|---|---|
-| **Start with your most-used docs** | Highest ROI knowledge sources | FAQ doc, product guide, pricing sheet |
-| **Use projects for dynamic data** | Always up-to-date, no re-uploading | Customer list project, inventory project |
-| **Create manual sources for gaps** | Fill knowledge holes quickly | Edge case handling, tone guidelines |
-| **Organize by topic** | Helps agent find relevant info faster | Separate "Product" and "Pricing" sources |
-| **Update regularly** | Stale knowledge = wrong answers | Quarterly review of all knowledge sources |
-| **Test with real questions** | Verify agent uses knowledge correctly | Ask questions you know the answers to |
-| **Use backlinks to audit** | Understand cross-dependencies | Check before editing shared projects |
-| **Upload diverse formats** | Different formats capture different knowledge | PDFs for detail, CSVs for data, URLs for current info |
+| **Start with the documents that you use most** | They answer the most questions | FAQ document, product guide, price sheet |
+| **Use projects for data that changes** | The agent always has the latest data | Customer list project, inventory project |
+| **Write manual sources for gaps** | You add a missing fact fast | Special cases, tone rules |
+| **Organize by topic** | The agent finds the correct fact faster | Separate "Product" and "Pricing" sources |
+| **Update regularly** | Old knowledge gives wrong answers | Review all knowledge sources each quarter |
+| **Do a test with real questions** | You see if the agent uses the knowledge correctly | Ask questions with answers that you know |
+| **Use backlinks for audits** | You see which agents depend on a project | Look at backlinks before you edit a shared project |
+| **Use different formats** | Each format holds a different type of fact | PDFs for detail, CSVs for data, links for current information |
---
## What's Next
-| Guide | What You'll Learn |
+| Guide | What you learn |
|---|---|
-| [Tools for AI Agents](./agent-tools.md) | Add action capabilities to your agents |
-| [Custom AI Agents](./ai-agents-getting-started.md) | Build and configure agents from scratch |
-| [Multi-Agents](./multi-agents.md) | Create teams of agents with shared knowledge |
-| [Projects & Databases](../workspaces/projects-databases.md) | Structure data for optimal agent training |
-| [Adding Genesis Context](../genesis/adding-context.md) | Upload context for Genesis app creation |
+| [Tools for AI Agents](./agent-tools.md) | Let your agents take actions |
+| [Custom AI Agents](./ai-agents-getting-started.md) | Build and configure agents from the start |
+| [Multi-Agents](./multi-agents.md) | Make teams of agents that share knowledge |
+| [Projects & Databases](../workspaces/projects-databases.md) | Organize data for agent knowledge |
+| [Adding Genesis Context](../genesis/adding-context.md) | Upload context to build a Taskade Genesis app |
---
diff --git a/genesis-living-system-builder/ai-features/agent-teams.md b/genesis-living-system-builder/ai-features/agent-teams.md
index 3ed0cfe..f8c8ce0 100644
--- a/genesis-living-system-builder/ai-features/agent-teams.md
+++ b/genesis-living-system-builder/ai-features/agent-teams.md
@@ -6,115 +6,133 @@ description: >-
# AI Agent Teams
-A single agent is powerful. A **team** of agents is transformative. AI Agent Teams let you group multiple specialized agents so they can work together — Taskade automatically routes each request to the agent best equipped to handle it.
+An AI Team puts several specialized agents in one chat. You send one message, and the mode that you pick decides which agents reply. In **Auto** mode, Taskade sends each message to the agent that fits it best.
{% hint style="info" %}
-**Think of it like a real team.** You don't ask your accountant to write marketing copy. Agent Teams apply the same principle: the right specialist handles the right task, every time.
+**An AI Team works like a real team.** You do not ask your accountant to write marketing copy. In an AI Team, the correct specialist does each task.
{% endhint %}
{% hint style="info" %}
-**Plan:** AI Teams need Pro or a higher plan (Pro, Business, Max, or Enterprise). Free and Starter workspaces can still use single custom agents. Each agent that replies uses AI credits, so an **Everyone** turn costs more than an **Auto** turn.
+**Plan:** AI Teams need Pro or a higher plan (Pro, Business, Max, or Enterprise). Free and Starter workspaces can still use single custom agents. Each agent that replies uses AI credits. Thus an **Everyone** turn costs more than an **Auto** turn.
{% endhint %}
## Create a Team
-1. Open the **AI Teams** tab in your workspace.
-2. Click **Create**.
+1. In your workspace sidebar, open **AI Teams**.
+2. Click **Create**. You can also click **+** (**Create AI team**) next to **AI Teams** and pick **Start Blank**.
3. In the **Create Team** dialog, type a **Name**.
-4. Under **Add to team**, add the agents you want (at least one). All agents in a team must be in the same workspace.
+4. Under **Add to team**, add the agents that you want. Add at least one agent.
5. Click **Create**.
-Your team is ready to use immediately.
+All agents in a team must be in the same workspace. You can use the team immediately.
+
+### Give Each Agent a Role
+
+A role is a short label, such as Researcher or Writer. It tells the team what each agent is for. In **Orchestrate** mode, the lead reads the roles to pick the agent for each step.
+
+1. In **AI Teams**, open the team.
+2. Click **Edit team**.
+3. Under **Manage team**, click **+ Add role** below the name of an agent.
+4. Type the role (up to 60 characters), and press **Enter**.
+5. Click **Update**.
+
+You can also ask Taskade EVE to rename a team, change its instruction, add or remove an agent, or set the role of each agent.
### Create a Team with AI
-Don't know which agents to include? Let AI figure it out.
+If you do not know which agents to include, let the AI Team Generator make them.
-1. In the **AI Teams** tab, click **Generate**.
-2. Describe what you need the team to do (e.g., "A content marketing team that handles blog writing, SEO optimization, and social media scheduling").
+1. In **AI Teams**, click **Generate**.
+2. Describe the work of the team (for example, "A content marketing team that handles blog writing, SEO optimization, and social media scheduling").
3. Press **Enter**.
-4. Taskade generates a fully configured team with specialized agents, roles, and tools.
-5. Review, adjust if needed, and save.
+
+Taskade creates the specialized agents, gives each one a role and tools, and groups them into a team. Review each agent, and change it if necessary.
{% hint style="success" %}
-**The AI Team Generator is the fastest way to get started.** Describe a business function and get a working team in seconds. You can always fine-tune individual agents later.
+**The AI Team Generator is the fastest way to start.** You describe a business function and get a team that works. You can change each agent later.
{% endhint %}
## How to Use Your Team
### Chat in the AI Teams Tab
-Go to **AI Teams** > select your team > start a conversation. In **Auto** mode, the best-suited agent replies. Switch modes to hear from more agents (see [Execution Modes](#execution-modes)).
+1. Open **AI Teams**.
+2. Select your team.
+3. Pick a mode above the message box.
+4. Type your message.
+
+In **Auto** mode, the best agent for the message replies. To hear from more agents, select a different mode (see [Execution Modes](#execution-modes)).
### Chat Inside a Project
-1. Open any project.
+1. Open a project.
2. Click the **chat balloon** icon in the top bar.
-3. Open the menu at the top of the chat and pick your team under the teams list.
-4. Start chatting.
+3. Open the menu at the top of the chat.
+4. Pick your team from the teams list.
+5. Type your message.
-The team responds using the project as context, so answers are grounded in your actual work.
+The team uses the project as context. Thus the answers come from your real work.
### Assign a Team to Tasks
-In any project, type **/** followed by the team name to assign the team to a task. The team picks the right agent to handle the work based on the task content.
+In a project, type **/** and the team name to assign the team to a task. The team picks the correct agent for the content of the task.
## Execution Modes
-Execution modes control **how** your team responds. Choose the mode that fits your workflow.
+The execution mode controls **how** your team replies. Choose the mode that fits your work.
| Mode | What Happens | Best For |
|---|---|---|
-| **Auto** | Taskade picks the single best agent for your message | Most everyday requests — fast and hands-off |
-| **Everyone** | Every agent on the team replies, one after another | Getting diverse perspectives or second opinions |
-| **Manual** | Only the agents you select reply, one after another | A fixed pipeline, such as researcher then writer |
-| **Orchestrate** | A lead plans the steps, gives each step to an agent, and adds a review step | Complex tasks that need multiple areas of expertise combined |
+| **Auto** | Taskade picks the single best agent for your message | Most everyday requests, fast and with no setup |
+| **Everyone** | Every agent on the team replies, one after another | Many points of view or second opinions |
+| **Manual** | Only the agents you select reply, one after another | A fixed sequence, such as researcher then writer |
+| **Orchestrate** | A lead plans the steps, gives each step to an agent, and adds a review step | Large tasks that need the skills of many specialists |
{% hint style="info" %}
-**Start with Auto mode.** It works well for most situations. Switch to other modes when you need more control or want to see how different agents approach the same problem.
+**Start with Auto mode.** Auto mode is correct for most requests. Other modes give you more control, or show how different agents do the same task.
{% endhint %}
### When to use each mode
-**Auto** — "Handle this customer inquiry" (let the system route it)
+**Auto**: "Handle this customer inquiry" (Taskade sends it to the correct agent)
-**Everyone** — "What should our Q2 strategy be?" (get input from every perspective)
+**Everyone**: "What should our Q2 strategy be?" (each agent gives its point of view)
-**Manual** — "Turn these notes into a blog post" with the Researcher and the Writer selected (only the agents you pick)
+**Manual**: "Turn these notes into a blog post", with the Researcher and the Writer selected (only the agents that you pick)
-**Orchestrate** — "Write a product launch plan covering marketing, pricing, and support" (multiple specialists, one cohesive output)
+**Orchestrate**: "Write a product launch plan covering marketing, pricing, and support" (many specialists, one final result)
## Use Cases
### Customer Support Team
-Combine agents specialized in billing, technical issues, and account management. Incoming questions automatically go to the right specialist.
+Put agents for billing, technical problems, and account management in one team. Each question goes to the correct specialist.
### Research Panel
-Group a market research agent, a competitive analysis agent, and a data analyst agent. Ask a strategic question and get a comprehensive answer from multiple angles.
+Put a market research agent, a competitor analysis agent, and a data analyst agent in one team. Ask a strategic question and get an answer from many points of view.
-### Content Creation Pipeline
+### Content Pipeline
-Team up a research agent, a writing agent, an SEO agent, and a social media agent. One prompt can produce a blog post, optimize it for search, and draft social posts to promote it.
+Put a research agent, a writing agent, an SEO agent, and a social media agent in one team. One prompt can make a blog post, improve it for search, and draft social posts for it.
### Sales Qualification
-Combine a lead research agent, a qualification agent, and an outreach agent. Feed in a list of prospects and get qualified leads with personalized follow-up drafts.
+Put a lead research agent, a qualification agent, and an outreach agent in one team. Give the team a list of prospects. You get qualified leads and personal follow-up drafts.
## Tips for Effective Teams
-* **Give each agent a clear specialty.** Overlapping roles cause confusion about who should respond. If two agents cover "marketing," narrow one to "content" and the other to "paid ads."
-* **Keep teams focused.** 3 to 5 agents per team is the sweet spot. Larger teams slow down routing and produce noisier results.
-* **Train agents individually first.** Make sure each agent performs well on its own before adding it to a team. A team is only as strong as its weakest member.
-* **Use descriptive team names.** "Customer Success Team" is better than "Team 1." Clear names help you (and your teammates) pick the right team quickly.
+* **Give each agent a clear specialty.** If two roles overlap, the team cannot decide which agent replies. If two agents cover "marketing," give one "content" and the other "paid ads."
+* **Keep teams small.** A team of 3 to 5 agents works best. Larger teams are slower and give less focused results.
+* **Do a test of each agent first.** Make sure that each agent works well alone before you add it to a team.
+* **Use clear team names.** "Customer Success Team" is better than "Team 1." Clear names help you and your teammates pick the correct team.
## Next Steps
-* [AI Agent Tools](agent-tools.md) — give each agent the ability to take action
-* [Agent Knowledge & Memory](agent-knowledge.md) — train agents with your business data
-* [AI Agents Getting Started](ai-agents-getting-started.md) — create individual agents before building teams
-* [Automation Recipes](../automation/recipes.md) — connect teams to automated workflows
+* [AI Agent Tools](agent-tools.md): let each agent take action.
+* [Agent Knowledge & Memory](agent-knowledge.md): give agents your business data.
+* [AI Agents Getting Started](ai-agents-getting-started.md): create single agents before you build teams.
+* [Automation Recipes](../automation/recipes.md): connect teams to automations.
---
diff --git a/genesis-living-system-builder/ai-features/agent-tools.md b/genesis-living-system-builder/ai-features/agent-tools.md
index 7361472..0a0f73c 100644
--- a/genesis-living-system-builder/ai-features/agent-tools.md
+++ b/genesis-living-system-builder/ai-features/agent-tools.md
@@ -9,20 +9,20 @@ description: >-
## Overview
-Agent Tools are integrated add-ons that **augment your AI agents' capabilities** — allowing them to interact with Taskade projects, third-party platforms, and custom workflows. Tools transform agents from text-only assistants into autonomous actors that can take real actions on your behalf.
+With tools, your AI agents can **act, not only chat**. An agent with tools can work in your Taskade projects, use other apps, and run your automations. You turn on each tool, and the agent uses it in the chat.
-> **Workspace DNA connection:** Tools are how agents execute the [Execution pillar](../automation/README.md) of Workspace DNA. An agent with tools can send Slack messages, create HubSpot deals, update Google Sheets, and more — all within the conversation.
+> **Workspace DNA connection:** Tools connect agents to the [Execution pillar](../automation/README.md) of Workspace DNA. An agent with tools can send Slack messages, create HubSpot deals, update Google Sheets, and more, inside the chat.
---
## Add Tools to Agents
-1. Go to the **Agents** tab
-2. Click the three dots (⋯) next to your agent → **Edit agent**
-3. Open the **Tools** tab
-4. Click **Add tools**
-5. Select from Taskade tools and/or third-party integrations
-6. Save
+1. Open **AI Agents** in your workspace sidebar.
+2. Click the three dots (⋯) next to your agent.
+3. Choose **Add tools**. The **Actions** tab of the agent opens on the **Tools** sub-tab.
+4. Turn on the Taskade tools and the third-party apps that the agent needs. Taskade saves each change.
+
+To filter the list, click **All**, **Active**, **Taskade**, **Apps**, or **Custom**.
@@ -30,15 +30,15 @@ Agent Tools are integrated add-ons that **augment your AI agents' capabilities**
## Use Agent Tools
-Agents can use tools through three interfaces:
+Agents can use tools in three places:
| Method | How | Best For |
|---|---|---|
-| **Project Chat** | Chat with an agent in a project context — agent uses tools automatically | Project-specific actions |
-| **Agent Sidebar** | Access agent in the sidebar — tools available in conversation | Quick actions while working |
-| **Agent Chat** | Dedicated agent chat session — full tool access | Complex multi-step tasks |
+| **Project Chat** | Chat with an agent inside a project. The agent uses its tools when the request needs them | Actions for one project |
+| **Agent Sidebar** | Open the agent in the sidebar. Its tools work in the chat | Quick actions while you work |
+| **Agent Chat** | Open a separate chat with the agent. All its tools work | Large tasks with many steps |
-> **Agents decide when to use tools.** You don't need to explicitly request a tool — describe what you want, and the agent determines which tool to use.
+> **The agent decides when to use a tool.** You do not name the tool. Describe what you want, and the agent picks the tool.
---
@@ -46,125 +46,129 @@ Agents can use tools through three interfaces:
### Agent Actions (Full Reference)
-These are built-in Taskade actions every agent can perform in your workspace:
+Every agent can do these built-in Taskade actions in your workspace:
| Category | Actions | Example Prompts |
|---|---|---|
-| **Bulk AI Prompts** | Run AI commands across multiple items | "Summarize all tasks in this project", "Translate all items to Spanish" |
-| **Project & Task Editing** | Create, update, delete projects and tasks | "Create a new task called 'Review proposal'", "Rename this project to Q2 Plan" |
-| **Content Updates** | Modify task descriptions, notes, content | "Add detailed instructions to the 'Setup server' task" |
-| **Task Completion** | Mark tasks complete or reopen them | "Mark all tasks under 'Phase 1' as complete" |
-| **Due Date Management** | Set, update, remove due dates | "Set the deadline for 'Launch campaign' to next Friday" |
+| **Bulk AI Prompts** | Run AI commands on many items at the same time | "Summarize all tasks in this project", "Translate all items to Spanish" |
+| **Project & Task Editing** | Create, update, and delete projects and tasks | "Create a new task called 'Review proposal'", "Rename this project to Q2 Plan" |
+| **Content Updates** | Change task descriptions, notes, and content | "Add detailed instructions to the 'Setup server' task" |
+| **Task Completion** | Mark tasks complete or open them again | "Mark all tasks under 'Phase 1' as complete" |
+| **Due Date Management** | Set, update, and remove due dates | "Set the deadline for 'Launch campaign' to next Friday" |
| **Assignments** | Assign or unassign team members | "Assign @Sarah to all tasks in the Marketing section" |
-| **Labels & Tags** | Add, remove, manage labels | "Add the 'urgent' label to all overdue tasks" |
-| **Task Movement** | Reorder, move tasks between sections | "Move all completed tasks to the 'Done' section" |
-| **Smart Editing** | Restructure, reorganize, reformat content | "Convert this brainstorm list into an organized project plan" |
-| **Notes & Comments** | Add notes, comments, annotations | "Add a note to this task explaining the delay" |
-| **Insights & Reporting** | Analyze project data, generate reports | "Give me a status summary of all projects due this week" |
-| **Governance & Safety** | Permission checks, access validation | "Who has access to this project?" |
-| **Custom Fields** | Create, update, manage custom fields | "Add a 'Priority' custom field with options: High, Medium, Low" |
-| **Search & Intelligence** | Search across workspace, find patterns | "Find all tasks assigned to @John that are overdue" |
-| **Content Generation** | Generate text, documents, summaries | "Write a project brief for the new product launch" |
-| **Compare & Review** | Diff content, review changes | "Compare this week's metrics to last week's" |
+| **Labels & Tags** | Add, remove, and manage labels | "Add the 'urgent' label to all overdue tasks" |
+| **Task Movement** | Reorder tasks and move them between sections | "Move all completed tasks to the 'Done' section" |
+| **Smart Editing** | Restructure and reformat content | "Convert this brainstorm list into an organized project plan" |
+| **Notes & Comments** | Add notes and comments | "Add a note to this task explaining the delay" |
+| **Insights & Reporting** | Analyze project data and make reports | "Give me a status summary of all projects due this week" |
+| **Governance & Safety** | Show access and permissions | "Who has access to this project?" |
+| **Custom Fields** | Create, update, and manage custom fields | "Add a 'Priority' custom field with options: High, Medium, Low" |
+| **Search & Intelligence** | Search the workspace and find patterns | "Find all tasks assigned to @John that are overdue" |
+| **Content Generation** | Write text, documents, and summaries | "Write a project brief for the new product launch" |
+| **Compare & Review** | Compare content and review changes | "Compare this week's metrics to last week's" |
### Taskade AI (Cross-Agent)
| Capability | What It Does | Example |
|---|---|---|
-| **Ask other agents** | Query another agent from within a conversation | "Ask the Sales Agent about our latest lead metrics" |
-| **Run agent commands** | Execute another agent's custom commands | "Run the Report Agent's `/weekly-summary` command" |
+| **Ask an agent** | Ask another agent a question inside a chat | "Ask the Sales Agent about our latest lead metrics" |
+| **Run an agent command** | Run a custom command of another agent | "Run the Report Agent's `/weekly-summary` command" |
### Media Tools
| Tool | What It Does | Example |
|---|---|---|
-| **Search Web** | Search the internet for current information | "Search for the latest pricing trends in SaaS" |
-| **Scrape Webpage** | Extract content from any web page | "Scrape the competitor's pricing page and summarize" |
-| **Transcribe YouTube** | Get full transcript from any YouTube video | "Transcribe this product demo video and create action items" |
+| **Search the web** | Search the internet for current information | "Search for the latest pricing trends in SaaS" |
+| **Read a web page** | Get the text content of a web page. **Summarize a web page** gives a short summary instead | "Read this pricing page and summarize the plans" |
+| **Transcribe a YouTube video** | Get the full transcript of a YouTube video | "Transcribe this product demo video and create action items" |
+
+> **Web search in public chats:** In your own chats, a custom agent can search the web and read a web page with no tool turned on. A public chat searches the web only when the agent has **Search the web** turned on. It reads a page only when **Read a web page** is on. In both cases, the **Media** row on the **Tool Permissions** tab of **Publish** must read **Visitors can use this**.
---
## App Tools (Third-Party Integrations)
-Connect your agents to external platforms for real-world actions:
+Connect your agents to other apps to do real work:
| Category | Platform | Capabilities | Example Use |
|---|---|---|---|
| **Communication** | **Slack** | Send channel messages, send DMs, create channels, find users | Agent posts daily standup to #team |
| | **Discord** | Send messages to channels and users | Agent posts community updates |
-| | **Microsoft Teams** | Create channels, send channel/chat messages | Agent notifies team of events |
-| | **WhatsApp Business** | Send messages to contacts | Agent handles customer comms |
-| | **Telegram** | Send messages, manage groups | Agent broadcasts announcements |
+| | **Microsoft Teams** | Create channels, send channel and chat messages | Agent tells the team about events |
+| | **WhatsApp Business** | Send messages to contacts | Agent sends customer messages |
+| | **Telegram** | Send messages, manage groups | Agent sends announcements |
| | **Twilio** | Send SMS messages | Agent sends appointment reminders |
| **Email** | **Gmail** | Send emails, find emails, create drafts | Agent sends follow-up emails |
-| | **Outlook** | Send emails, manage calendar | Agent handles enterprise email |
+| | **Outlook** | Send emails, manage calendar | Agent sends company email |
| **CRM & Sales** | **HubSpot** | Create contacts, create deals, update records | Agent creates CRM records |
-| | **Salesforce** | Manage contacts, opportunities, cases | Agent updates enterprise CRM |
-| | **Apollo** | Enrich contacts, find leads | Agent enriches lead data |
-| **Productivity** | **Google Sheets** | Insert rows, find rows, update cells | Agent logs data to sheets |
-| | **Google Docs** | Create documents, append text, find documents | Agent generates documents |
+| | **Salesforce** | Manage contacts, opportunities, cases | Agent updates the company CRM |
+| | **Apollo** | Enrich contacts, find leads | Agent adds data to leads |
+| **Productivity** | **Google Sheets** | Add rows, find rows, update cells | Agent logs data to sheets |
+| | **Google Docs** | Create documents, add text, find documents | Agent writes documents |
| | **Google Calendar** | Create events, create quick events | Agent schedules meetings |
-| | **Google Drive** | Create folders/files, search, duplicate, manage permissions | Agent organizes files |
-| | **Google Forms** | Process form responses | Agent handles form data |
+| | **Google Drive** | Create folders and files, search, duplicate, manage permissions | Agent organizes files |
+| | **Google Forms** | Process form responses | Agent processes form data |
| | **Calendly** | Manage scheduling | Agent books appointments |
-| | **Typeform** | Process form submissions | Agent handles surveys |
+| | **Typeform** | Process form submissions | Agent processes surveys |
| **Social Media** | **Twitter/X** | Create posts, manage content | Agent posts social updates |
| | **LinkedIn** | Create company updates, publish articles | Agent publishes company news |
| | **Facebook** | Create page posts, manage content | Agent manages page posts |
-| | **Reddit** | Create posts, monitor threads | Agent engages communities |
+| | **Reddit** | Create posts, monitor threads | Agent takes part in communities |
| | **YouTube** | Transcribe videos, manage content | Agent processes video content |
| **Content** | **WordPress** | Create blog posts, manage content | Agent publishes content |
| | **Webflow** | Manage CMS content | Agent updates web content |
-| | **RSS** | Monitor feeds | Agent tracks content sources |
-| **Development** | **GitHub** | Manage issues, PRs, repositories | Agent tracks development |
-| | **HTTP/Webhooks** | Custom API requests with analyzer | Agent connects to any API |
+| | **RSS** | Monitor feeds | Agent follows content sources |
+| **Development** | **GitHub** | Manage issues, PRs, repositories | Agent follows development work |
+| | **HTTP Request** | Call a REST API from a custom tool automation | Agent connects to any API |
| **Business** | **Stripe** | Process payments, manage subscriptions | Agent handles billing |
| | **Mailchimp** | Add contacts, manage campaigns | Agent manages email marketing |
-> Agents can use any automation flow as a custom tool, and connect to 100+ integrations. On the Free plan, premium apps show an upgrade chip. Premium integrations need a paid plan, and Pro integrations need Pro and up. See [Automations](../automation/README.md) for the full integration list.
+> An agent can use an automation as a custom tool, and it can connect to 100+ integrations. On the Free plan, premium apps show an upgrade badge. Premium integrations need a paid plan, and Pro integrations need Pro and up. See [Automations](../automation/README.md) for the full list of integrations.
-> **For developers:** The **HTTP/Webhooks** custom tool lets agents call any external API. See [Webhooks](../../apis-living-system-development/webhooks.md) for programmatic event delivery and payload details.
+> **For developers:** To let an agent call an external API, add an **HTTP Request** step to a custom tool automation. See [Webhooks](../../apis-living-system-development/webhooks.md) for event delivery and payload details.
---
## Build Custom Agent Tools
-Turn any automation flow into a reusable agent tool:
+You can make a reusable agent tool from an automation:
+
+1. Open the **Actions** tab of your agent, then the **Tools** sub-tab.
+2. Click **New tool**.
+3. Define the parameters of the tool.
+4. Add the steps of the automation.
+5. Turn on the automation.
-1. Open your agent's **Tools** tab
-2. Click **New tool**
-3. Design the automation flow (trigger → actions)
-4. Save as a custom tool
-5. The agent can now execute this flow on command
+The automation shows as a new tool of the agent. Ask the agent to use the tool in the chat, or let the agent decide.
**Example custom tools:**
-| Custom Tool | Automation Flow | Trigger |
+| Custom Tool | Automation Steps | Trigger |
|---|---|---|
-| "Generate weekly report" | Gather project data → Summarize → Email to team → Post to Slack | Agent decides or `/weekly-report` command |
-| "Onboard new client" | Create project → Add template tasks → Send welcome email → Notify team | Agent processes new client request |
-| "Escalate issue" | Create urgent ticket → Notify manager via Slack → Schedule follow-up | Agent detects high-severity issue |
+| "Generate weekly report" | Collect project data → Summarize → Email to team → Post to Slack | The agent decides, or the `/weekly-report` command |
+| "Onboard new client" | Create project → Add template tasks → Send welcome email → Tell the team | The agent gets a request for a new client |
+| "Escalate issue" | Create urgent ticket → Tell the manager in Slack → Schedule follow-up | The agent finds a high-severity problem |
-> **Deep dive:** [Automations: The Execution Pillar](../automation/README.md) for full automation flow building.
+> **More detail:** Read [Automations: The Execution Pillar](../automation/README.md) to build automations.
---
## Human in the Loop
-For sensitive external actions, enable approval before execution:
+For actions in other apps, you can approve each action before it runs:
| Feature | Detail |
|---|---|
-| **What it is** | An approve/reject gate before agents interact with external platforms |
-| **When to use** | Sending emails, posting to social media, creating CRM records, any external action |
-| **How it works** | Agent proposes the action → You review → Approve or Reject → Agent proceeds or stops |
+| **What it is** | A gate where you approve or reject an action before the agent runs it |
+| **When to use** | Emails, social media posts, CRM records, and other external actions |
+| **How it works** | The agent proposes the action → You review it → You approve or reject → The agent continues or stops |
-1. Open the agent's **Tools** tab
-2. Set the approval mode of each sensitive tool to **Manual Approval** (routine tools can use **Automatic Approval**)
-3. When the agent wants to use a tool set to Manual Approval, you'll see a confirmation prompt
-4. Click **Approve** to proceed or **Reject** to stop
+1. Open the **Actions** tab of the agent, then the **Tools** sub-tab.
+2. Set the approval mode of each sensitive tool to **Manual Approval**. Routine tools can stay on **Automatic Approval**.
+3. Ask the agent to do a task. When the agent wants to use a **Manual Approval** tool, it pauses and asks you.
+4. Click **Approve** to continue, or **Reject** to stop.
-> **Tip:** Use Human in the Loop during the first few weeks of agent deployment. Once you trust the agent's judgment, switch routine tools to **Automatic Approval**.
+> **Tip:** Use **Manual Approval** for the first weeks with a new agent. When you trust the agent, set routine tools to **Automatic Approval**.
---
@@ -172,22 +176,22 @@ For sensitive external actions, enable approval before execution:
| Action | How |
|---|---|
-| Enable a tool | Toggle ON in the Tools tab |
-| Disable a tool | Toggle OFF in the Tools tab |
+| Enable a tool | Toggle ON in the **Tools** sub-tab of the **Actions** tab |
+| Disable a tool | Toggle OFF in the **Tools** sub-tab of the **Actions** tab |
| Remove a tool | Click the remove button next to the tool |
-> **Disabled tools** remain configured but are not available to the agent. Re-enable anytime without reconfiguration.
+> **A disabled tool** keeps its settings, but the agent cannot use it. You can turn it on again at any time without new settings.
---
## What's Next
-| Guide | What You'll Learn |
+| Guide | What you learn |
|---|---|
-| [Multi-Agents](./multi-agents.md) | Build teams of agents with complementary tools |
-| [Agent Knowledge & Memory](./agent-knowledge.md) | Train agents on your data |
-| [Automations](../automation/README.md) | Build the automation flows agents can trigger |
-| [Custom AI Agents](./ai-agents-getting-started.md) | Complete agent setup guide |
+| [Multi-Agents](./multi-agents.md) | Build teams of agents with different tools |
+| [Agent Knowledge & Memory](./agent-knowledge.md) | Give agents your data |
+| [Automations](../automation/README.md) | Build the automations that agents can start |
+| [Custom AI Agents](./ai-agents-getting-started.md) | Full guide to agent setup |
---
diff --git a/genesis-living-system-builder/ai-features/ai-agents-getting-started.md b/genesis-living-system-builder/ai-features/ai-agents-getting-started.md
index 8890878..cab90d6 100644
--- a/genesis-living-system-builder/ai-features/ai-agents-getting-started.md
+++ b/genesis-living-system-builder/ai-features/ai-agents-getting-started.md
@@ -5,39 +5,39 @@ description: >-
# AI Agents Getting Started
-Transform your workspace with smart AI assistants that understand your business and can handle tasks automatically!
+An AI agent is an assistant that knows your business. You can give it your documents and projects. Then it answers questions and does tasks for you.
## What are AI Agents?
-AI Agents are specialized assistants designed to automate activities like research, data analysis, and content creation. Learn more about [Taskade AI Agents](https://www.taskade.com/agents) and what they can do for your team. Think of them as smart team members who:
+An AI agent is an assistant with one job, such as research, data analysis, or content creation. Read more about [Taskade AI Agents](https://www.taskade.com/agents). An agent is like a team member that can do these things:
-* **Learn your business** by reading your documents and projects
-* **Answer questions** about your work and processes
-* **Take actions** like creating tasks, sending emails, or updating spreadsheets
-* **Work 24/7** without breaks, weekends, or vacations
-* **Specialize in specific roles** like marketing, legal, customer service, or research
+* **Learn your business**: It reads your documents and projects.
+* **Answer questions**: It answers questions about your work and processes.
+* **Take actions**: It creates tasks, sends emails, or updates spreadsheets.
+* **Work at all hours**: It works 24/7, also on weekends.
+* **Do one role**: It works as a marketing, legal, customer service, or research specialist.
-**Real example:** Your "Customer Service Agent" can instantly answer questions about your products, create support tickets, and even draft response emails - all based on your existing knowledge base.
+**Example:** A "Customer Service Agent" answers questions about your products. It also creates support tickets and drafts reply emails. It uses your existing knowledge base for all of these tasks.
### What Can AI Agents Help With?
| **Area** | **Tasks** |
| -------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
-| **Task Automation** | • Auto-categorize tasks
• Prioritize by keywords
• Estimate duration
• Recommend task owners
|
-| **Knowledge Retrieval** | • Fetch relevant documents
• Answer team questions
• Recommend reading material
• Suggest internal resources
|
-| **Feedback & Improvement** | • Collect and analyze feedback
• Offer improvement insights
• Benchmark against past projects
• Highlight success stories
|
-| **Meetings** | • Summarize discussions
• Recommend next steps
• Extract action items
|
-| **Creative Assistance** | • Offer writing/editing aid
• Generate content ideas
• Review and improve copy
|
+| **Task Automation** | • Sort tasks into categories
• Prioritize by keywords
• Estimate duration
• Recommend task owners
|
+| **Knowledge Retrieval** | • Find related documents
• Answer team questions
• Recommend reading material
• Suggest internal resources
|
+| **Feedback & Improvement** | • Collect and analyze feedback
• Suggest improvements
• Compare with past projects
• Show success stories
|
+| **Meetings** | • Summarize discussions
• Recommend next steps
• Find action items
|
+| **Creative Assistance** | • Help with writing and editing
• Generate content ideas
• Review and improve copy
|
-## When Should You Create an AI Agent?
+## When to Create an AI Agent
-Create an agent if you:
+An agent is a good choice when one of these conditions is true:
-* **Repeat the same tasks** multiple times per week
-* **Spend 2+ hours weekly** on similar activities
-* **Answer the same questions** over and over
-* **Need help** with research, writing, or analysis
-* **Want to streamline** team processes
+* You do the same tasks many times each week.
+* You spend 2 or more hours each week on similar work.
+* You answer the same questions again and again.
+* You want help with research, writing, or analysis.
+* You want to make team processes faster.
## Industry-Specific AI Agent Examples
@@ -46,16 +46,16 @@ Create an agent if you:
**Custom AI Agents for Legal Practice:**
{% hint style="success" %}
-Legal AI agents can automate document review, conduct legal research, and manage cases efficiently, saving lawyers significant time on routine tasks while maintaining accuracy and compliance.
+A legal agent can review documents, do legal research, and help manage cases. Lawyers then spend less time on routine work.
{% endhint %}
**Legal Agent Capabilities:**
-* **Document Review**: Analyze contracts, briefs, and legal documents for key clauses and potential issues
-* **Legal Research**: Search through case law, statutes, and legal precedents
-* **Case Management**: Track deadlines, court dates, and document filing requirements
-* **Client Communication**: Draft responses to common legal inquiries
-* **Compliance Monitoring**: Stay updated on regulatory changes and requirements
+* **Document Review**: Find key clauses and possible problems in contracts, briefs, and other legal documents.
+* **Legal Research**: Search case law, statutes, and legal precedents.
+* **Case Management**: Track deadlines, court dates, and filing requirements.
+* **Client Communication**: Draft replies to common legal questions.
+* **Compliance Monitoring**: Track changes to regulations and requirements.
**Training Your Legal Agent:**
@@ -66,13 +66,13 @@ Legal AI agents can automate document review, conduct legal research, and manage
* Statutes, regulations, and legislative history
* Citation guidelines and legal journal articles
* Argumentation and negotiation materials
-* Precedent-setting case laws and historical documents
+* Precedent cases and historical documents
* Legal textbooks, treatises, and specialized databases
**Legal Agent Templates:**
* **👩⚖️ Legal Advisor**: General legal guidance and document analysis
-* **📋 Contract Reviewer**: Specialized in contract analysis and risk assessment
+* **📋 Contract Reviewer**: Contract analysis and risk assessment
* **🔍 Research Assistant**: Legal research and case law analysis
* **⚖️ Compliance Monitor**: Regulatory compliance and updates
@@ -82,10 +82,10 @@ Legal AI agents can automate document review, conduct legal research, and manage
**Healthcare Agent Capabilities:**
-* **Patient Documentation**: Assist with medical record organization and analysis
-* **Research Support**: Find latest medical research and treatment protocols
-* **Compliance Assistance**: Help maintain HIPAA and regulatory compliance
-* **Scheduling Optimization**: Coordinate appointments and resource allocation
+* **Patient Documentation**: Help organize and analyze medical records.
+* **Research Support**: Find recent medical research and treatment protocols.
+* **Compliance Assistance**: Help with HIPAA and other regulatory requirements.
+* **Scheduling**: Coordinate appointments and resources.
### Marketing Teams
@@ -93,10 +93,10 @@ Legal AI agents can automate document review, conduct legal research, and manage
**Marketing Agent Capabilities:**
-* **Content Creation**: Generate blog posts, social media content, and marketing copy
-* **Campaign Analysis**: Analyze performance metrics and suggest optimizations
-* **Audience Research**: Identify target demographics and market trends
-* **Competitive Intelligence**: Monitor competitor activities and strategies
+* **Content Creation**: Write blog posts, social media content, and marketing copy.
+* **Campaign Analysis**: Analyze performance metrics and suggest improvements.
+* **Audience Research**: Find target audiences and market trends.
+* **Competitive Intelligence**: Track the activities and strategies of competitors.
### Customer Service Teams
@@ -104,58 +104,62 @@ Legal AI agents can automate document review, conduct legal research, and manage
**Customer Service Agent Capabilities:**
-* **Ticket Routing**: Automatically categorize and prioritize support requests
-* **Knowledge Base**: Provide instant answers to common customer questions
-* **Escalation Management**: Identify when to escalate issues to human agents
-* **Customer Feedback**: Analyze feedback patterns and suggest improvements
+* **Ticket Routing**: Sort and prioritize support requests.
+* **Knowledge Base**: Answer common customer questions at once.
+* **Escalation Management**: Find the requests that need a human.
+* **Customer Feedback**: Find patterns in feedback and suggest improvements.
## Create Your First AI Agent in 5 Steps
-### Step 1: Go to the Agents Tab
+### Step 1: Open AI Agents
-1. Open your Taskade workspace
-2. Click the **🤖 Agents** tab at the top
-3. Click **➕ Create agent**
+1. Open your Taskade workspace.
+2. In the workspace sidebar, open **🤖 AI Agents**.
+3. Click **Generate** to describe the agent, **Templates** to pick one, or **Create** to start blank.
+
+{% hint style="info" %}
+Every plan includes custom agents. Free includes one custom agent with one knowledge source. Pro and up have no agent limit.
+{% endhint %}
### Step 2: Choose How to Create Your Agent
-Pick the best option for you:
+Pick the option that fits your work:
-**🎯 Describe What You Need** (Recommended for beginners)
+**🎯 Describe What You Need** (best for beginners)
-* Simply tell Taskade what you want your assistant to do
-* We'll build it automatically for you
-* Works great for common business tasks
+* Click **Generate** and tell Taskade what the agent must do.
+* Taskade builds the agent for you.
+* This option works well for common business tasks.
**🛠️ Build It Yourself**
-* Create a custom assistant with specific abilities
-* You decide exactly how it behaves and responds
-* Best when you have unique requirements
+* Click **Create** to start with a blank agent.
+* You decide how the agent behaves and replies.
+* This option is best for special requirements.
**📋 Pick from Ready-Made Templates**
-* Choose from assistants for Marketing, Support, HR, and more
-* They're ready to work immediately
-* You can customize them anytime
+* Click **Templates** and pick an agent for Marketing, Support, HR, or another role.
+* A template agent is ready to use at once.
+* You can change it at any time.
**👥 Use Community Templates**
-* Browse assistants created by other businesses
-* Copy the ones that work for your needs
-* Learn from successful setups
+* Browse agents that other businesses made.
+* Copy the agents that fit your needs.
+* Learn from setups that work.
-### Step 3: Tell Your Assistant What to Do
+### Step 3: Tell Your Agent What to Do
-Give your AI assistant a clear role and personality - just like hiring a new team member:
+Give your agent a clear role and personality. Do this as you do for a new team member.
-**Name:** Make it memorable and relevant\
+**Name:** Use a name that is easy to remember and relevant.\
_Examples: "Sarah - Customer Success", "Marketing Mike", "Data Dave"_
-**Description:** Explain what your agent does\
+**Description:** Write what your agent does.\
_Example: "You are a customer success specialist who helps clients get the most value from our product. You're friendly, knowledgeable, and always focused on solving problems."_
-**Tone:** Choose how your agent should communicate
+**Tone:** Pick how your agent writes:
* Friendly and casual
* Professional and formal
@@ -164,25 +168,25 @@ _Example: "You are a customer success specialist who helps clients get the most
### Step 4: Train Your Agent with Knowledge
-Make your agent smart by giving it information:
+Give your agent information to answer from. These are the source types:
**📄 Upload Documents**
-* Drag and drop PDFs, Word docs, spreadsheets
-* Include product manuals, processes, FAQs
-* Add training materials and guidelines
+* Drag PDFs, Word documents, and spreadsheets into the agent.
+* Add product manuals, processes, and FAQs.
+* Add training materials and guidelines.
**🌐 Add Websites**
* Your company website and blog
-* Competitor information
+* Information about competitors
* Industry resources and news
**📋 Connect Projects**
-* Link existing Taskade projects as knowledge
-* Agent learns from your current work
-* Stays updated as projects change
+* Link existing Taskade projects as knowledge.
+* The agent learns from your current work.
+* The agent gets each change when a project changes.
**🎥 YouTube Videos**
@@ -190,89 +194,94 @@ Make your agent smart by giving it information:
* Product demos and presentations
* Industry webinars and conferences
+Website links, YouTube videos, and cloud files (Google Drive, Dropbox, Box, OneDrive) need a paid plan. On Free, each agent holds one knowledge source. Paid plans hold up to 50 sources for each agent.
+
### Step 4.5: Choose Your AI Model
-Taskade's model picker is powered by **TSK-1**, the intelligence layer that auto-routes every request to the right capability tier. You pick what you need the agent to do; TSK-1 handles which model executes it:
+The model picker uses **TSK-1**. TSK-1 is the Taskade system that sends each request to the correct capability tier. You pick what the agent must do. TSK-1 picks the model that does it.
| Tier | When to use it |
|------|---------------|
-| **Auto** (recommended) | TSK-1 picks the right tier per request — fast when you need speed, deep when you need reasoning. New agents default to Auto on every plan |
+| **Auto** (recommended) | TSK-1 picks the correct tier for each request. It is fast for simple work and deep for hard reasoning. New agents use Auto on every plan |
| **Instant** | Fast edits, simple completions, quick lookups |
-| **Standard** | Everyday work — balanced speed and reasoning |
+| **Standard** | Everyday work, with a balance of speed and reasoning |
| **Thinking** | Planning, creation, multi-step reasoning |
| **Reasoning** | Complex systems, deep analysis, multi-agent coordination |
{% hint style="info" %}
-**Vendor-agnostic.** TSK-1 runs on the latest frontier models. The tier describes what the system is good at, not which provider powers it — providers may change. Auto and Instant work on every plan. Standard and Thinking need a paid plan, and Reasoning needs Business or a higher plan. Agents pinned to a specific model (each model shows its plan badge in the picker) bypass TSK-1 routing. See [TSK-1 Model Tiers](tsk-1-model-tiers.md) for the full reference.
+**Vendor-agnostic.** TSK-1 runs on frontier models from top AI labs. A tier describes what the system does well, not which provider runs it. Providers can change. Auto and Instant work on every plan. Standard and Thinking need a paid plan, and Reasoning needs Business or a higher plan. An agent set to one specific model does not use TSK-1 routing. Each model shows its plan badge in the picker. See [TSK-1 Model Tiers](tsk-1-model-tiers.md) for the full reference.
{% endhint %}
## Understanding Agent Knowledge & Memory
-Your agent's knowledge is what makes it truly powerful for your business. This is the foundation that powers Taskade Genesis apps - your workspace's living DNA.
+The knowledge of your agent decides how useful it is for your business. The same knowledge also feeds your Taskade Genesis apps. Taskade calls this shared base the Workspace DNA of your workspace.
### Types of Agent Memory
#### **Session Memory**
-* Remembers conversation context during a chat
-* Maintains topic continuity
-* Resets when conversation ends
+* Keeps the context of the current chat
+* Keeps the chat on topic
+* Ends when the chat ends
#### **Knowledge Memory**
-* Permanent information from uploaded documents
-* Your business-specific data and processes
-* Forms the core intelligence for Genesis apps
+* Keeps the information from your uploaded documents
+* Holds your business data and processes
+* Gives your Taskade Genesis apps their core knowledge
+
+This knowledge is the persistent memory of the agent. It stays with the agent in every chat.
#### **Project Memory**
-* Live connection to your workspace projects
-* Updates automatically as your data changes
-* Powers the "living DNA" that makes each Genesis app unique
+* Keeps a live connection to your workspace projects
+* Updates when your data changes
+* Makes each Taskade Genesis app specific to your workspace
### Advanced Knowledge Training
#### **Bulk Training**
-Upload multiple knowledge sources at once:
+You can add many knowledge sources at the same time:
-* Multiple files simultaneously
-* Several web links in one batch
-* Connect multiple projects together
-* Perfect for comprehensive agent training
+* Many files in one upload
+* Many web links in one batch
+* Many projects in one selection
#### **Knowledge Formats Supported**
| **File Types** | **Web Resources** |
| ------------------------- | --------------------- |
-| 📄 .pdf, .docx, .txt, .md | 🎥 YouTube videos |
-| 📊 .xlsx, .csv | 📰 Blog articles |
-| 📋 .pptx | 🐦 Social media posts |
-| 📚 .epub | 📰 News websites |
+| 📄 .pdf, .docx, .txt, .md | 🌐 Website URLs |
+| 📊 .xlsx, .xls, .csv | 📰 Blog articles |
+| 📋 .pptx | 🎥 YouTube videos |
+| 📚 .epub | 📡 RSS / Atom feeds, sitemap XML |
+
+The Knowledge tab does not accept image files. Web resources need a paid plan.
#### **Knowledge Management Best Practices**
-* **Review sources** regularly for relevance
-* **Update documents** when business information changes
-* **Remove outdated** resources to maintain accuracy
-* **Test knowledge** by asking your agent questions
-* **Monitor performance** and refine training data
+* **Review sources** at regular intervals.
+* **Update documents** when your business information changes.
+* **Remove old** resources to keep answers correct.
+* **Test knowledge**: ask your agent questions.
+* **Monitor answers** and improve the sources.
### Step 5: Create Custom Commands
-Give your agent specific skills with custom commands:
+A custom command gives your agent a specific skill. You start a command when you type `/` and its name.
**Examples for a Sales Agent:**
-* `/qualify-lead` - Analyze if a prospect is worth pursuing
-* `/write-proposal` - Generate a custom proposal
-* `/follow-up` - Create follow-up email sequences
+* `/qualify-lead`: Find out if a prospect is worth the effort
+* `/write-proposal`: Write a custom proposal
+* `/follow-up`: Write a series of follow-up emails
**Examples for a Support Agent:**
-* `/troubleshoot` - Help diagnose customer problems
-* `/escalate` - Determine when to involve a human
-* `/knowledge-search` - Find relevant help articles
+* `/troubleshoot`: Find the cause of a customer problem
+* `/escalate`: Decide when a human must help
+* `/knowledge-search`: Find related help articles
## Popular Agent Types & Examples
@@ -284,13 +293,13 @@ Give your agent specific skills with custom commands:
### Customer Support Agent
-**What it does:** Answer questions, troubleshoot issues, route tickets\
+**What it does:** Answer questions, find the cause of problems, route tickets\
**Sample prompt:** "A customer can't log into their account. What should I tell them?"\
**Best for:** Support teams, customer success, help desks
### Data Analysis Agent
-**What it does:** Interpret data, create reports, find insights\
+**What it does:** Read data, create reports, find insights\
**Sample prompt:** "Analyze this month's sales data and highlight key trends"\
**Best for:** Analytics teams, managers, business owners
@@ -302,7 +311,7 @@ Give your agent specific skills with custom commands:
### Research Agent
-**What it does:** Gather information, summarize findings, competitive analysis\
+**What it does:** Collect information, summarize findings, analyze competitors\
**Sample prompt:** "Research the top 5 competitors in the project management space"\
**Best for:** Strategy teams, consultants, analysts
@@ -310,53 +319,53 @@ Give your agent specific skills with custom commands:
### Method 1: Agent Chat
-1. Go to the **Agents** tab
-2. Select your agent
-3. Click **➕ New chat**
-4. Start asking questions or giving tasks
+1. Open **AI Agents** in your workspace sidebar.
+2. Select your agent.
+3. On the **Messages** tab, click **➕** to start a new chat.
+4. Ask questions or give tasks.
### Method 2: Project Integration
-1. Open any project
-2. Click the chat icon (💬) at the top right
-3. Select your agent from the dropdown
-4. Get help with that specific project
+1. Open a project.
+2. Click the chat icon (💬) at the top right.
+3. Select your agent from the dropdown.
+4. Ask for help with that project.
### Method 3: Quick Commands
-1. In any project, type **/** followed by your command name
-2. Example: `/write-summary` or `/analyze-data`
-3. Press Enter to execute
+1. In a project, type **/** and then the name of your command.
+2. For example, type `/write-summary` or `/analyze-data`.
+3. Press Enter to run the command.
### Method 4: Agent Sidebar
-1. Click the **⊞ Manage agents** button
-2. Star your favorite agents
-3. Access them quickly from the right sidebar
+1. Click the **⊞ Manage agents** button.
+2. Star your favorite agents.
+3. Open them from the right sidebar.
## **Enhanced File Attachments**
-Taskade EVE conversations now support rich file interactions with complete visibility and context.
+You can attach files to a chat with Taskade EVE, the AI assistant of your workspace. The files show in the chat with your messages.
-### **Drag & Drop File Magic**
+### **Drag and Drop Files**
-Simply drag any file directly into your EVE chat:
+Drag a file into your Taskade EVE chat. Taskade EVE accepts these file types:
-* **Images**: JPG, PNG, GIF, WebP (up to 20 MB each) - EVE analyzes visual content
-* **Documents**: PDF (up to 20 MB each), TXT, MD - EVE reads and processes text
-* **Data**: CSV, JSON (text files up to 200 KB each) - EVE analyzes data and creates insights
-* **Code**: paste a snippet into the message - EVE reviews and suggests improvements
+* **Images**: JPG, PNG, GIF, WebP (up to 20 MB each). Taskade EVE analyzes the image.
+* **Documents**: PDF (up to 20 MB each), TXT, MD. Taskade EVE reads the text.
+* **Data**: CSV, JSON (text files up to 200 KB each). Taskade EVE analyzes the data.
+* **Code**: Paste a code snippet into the message. Taskade EVE reviews it and suggests improvements.
-For larger files or material the agent needs on every reply, add them to the agent's knowledge instead.
+If a file is larger, or the agent needs it in every reply, add the file to the knowledge of the agent.
### **Visible File Context**
-Files now appear directly alongside your messages instead of being hidden:
+Each file shows next to the message that it came with:
-* **Chronological Organization**: See exactly when each file was shared
-* **Visual Thumbnails**: Quick preview of images and documents
-* **File Status Indicators**: Clear confirmation when files are processed
-* **Conversation Flow**: Files integrate naturally with your discussion
+* **Chronological Organization**: You see when each file came into the chat.
+* **Visual Thumbnails**: You see a small preview of images and documents.
+* **File Status Indicators**: You see when Taskade finished with a file.
+* **Conversation Flow**: The files stay in the order of the chat.
### **Professional Use Cases**
@@ -390,126 +399,120 @@ EVE: "I've reviewed your Python code. Here are potential security concerns..."
### **File Management Features**
-* **Rename Conversations**: Turn "Chat with EVE" into "Q3 Report Analysis"
-* **File History**: Access previously shared files from conversation history
-* **Cross-Reference**: EVE remembers files from earlier in the conversation
-* **Batch Processing**: Upload multiple related files for comprehensive analysis
+* **Rename Conversations**: Change "Chat with EVE" to a clear name, such as "Q3 Report Analysis".
+* **File History**: Open files from earlier in the chat history.
+* **Cross-Reference**: Taskade EVE remembers files from earlier in the same chat.
+* **Batch Processing**: Upload many related files for one analysis.
## Advanced Features
### Agent Tools
-Give your agents superpowers by enabling tools:
+Tools let your agent do work, not only write text. Turn on the tools that the agent needs:
-**📝 Taskade Tools** - Create tasks, update projects, manage data\
-**🌐 Web Search** - Find real-time information online\
-**📧 Email Integration** - Send emails through Gmail\
-**💬 Communication** - Post to Slack, Teams, Discord\
-**📊 Data Tools** - Work with Google Sheets, databases\
-**🔄 Automation** - Trigger workflows and processes
+**📝 Taskade Tools**: Create tasks, update projects, and manage data\
+**🌐 Web Search**: Find current information on the web\
+**📧 Email Integration**: Send emails with Gmail\
+**💬 Communication**: Post to Slack, Teams, and Discord\
+**📊 Data Tools**: Work with Google Sheets and databases\
+**🔄 Automation**: Start automations and processes
### Multi-Agent Teams (Advanced)
{% hint style="success" %}
-Multi-Agent Teams put specialized AI agents in one chat, so each specialist handles its part of the work while you focus on strategic work. AI Teams need Pro or a higher plan.
+A Multi-Agent Team puts specialized AI agents in one chat. Each specialist does its part of the work. AI Teams need Pro or a higher plan.
{% endhint %}
#### What Are Multi-Agent Teams?
-Multi-agents are an evolution of Taskade's agentic workflow designed to:
+A multi-agent team is a group of agents that work together:
-* **Split the work** so each agent handles the part it knows best
-* **Collaborate intelligently** with each agent bringing specialized expertise
-* **Learn your preferences** and adapt to your communication style
-* **Maintain context** across complex, multi-step workflows
+* **Split the work**: Each agent does the part that it knows best.
+* **Collaborate**: Each agent adds its own specialty.
+* **Learn your preferences**: The agents adapt to your style of communication.
+* **Keep context**: The agents keep the context across work with many steps.
#### Setting Up Your Multi-Agent Team
**Method 1: Build from Existing Agents**
-1. Open **AI Teams** in your workspace sidebar
-2. Click **Create** and type a team name
-3. Under **Add to team**, add existing agents with complementary skills
-4. Click **Create**, then use **Edit team** to give each agent a role
+1. Open **AI Teams** in your workspace sidebar.
+2. Click **Create** and type a team name.
+3. Under **Add to team**, add existing agents with different skills.
+4. Click **Create**.
+5. Use **Edit team** to give each agent a role.
**Method 2: Create New Team from Templates**
-1. Go to **AI Teams** section
-2. Click **Templates**
-3. Choose from pre-built team configurations:
- * **Marketing Team**: Content + Social Media + Analytics agents
- * **Customer Success Team**: Support + Documentation + Training agents
- * **Product Development Team**: Research + Design + Testing agents
- * **Operations Team**: Logistics + Communication + Reporting agents
+1. Open **AI Teams**.
+2. Click **Templates**.
+3. Pick a team template. These are examples:
+ * **Marketing Team**: Content, Social Media, and Analytics agents
+ * **Customer Success Team**: Support, Documentation, and Training agents
+ * **Product Development Team**: Research, Design, and Testing agents
+ * **Operations Team**: Logistics, Communication, and Reporting agents
**Method 3: Custom Team Creation**
-1. Start with the **AI Team Generator** (see below)
-2. Describe your ideal team composition and goals
-3. Let AI automatically configure specialized roles and behaviors
-4. Fine-tune the team based on your specific requirements
+1. Start with the **AI Team Generator**. Read [AI Team Generator: Multi-Agent Teams](#ai-team-generator-multi-agent-teams) on this page.
+2. Describe the team and its goals.
+3. Let Taskade create the roles and instructions.
+4. Change the team to fit your requirements.
#### Using Multi-Agent Teams in Projects
**Option 1: Agent Sidebar Integration**
-* Access your multi-agent team through the persistent agent sidebar
-* Agents work in the background with real-time updates
-* Pre-trained on your knowledge base and third-party integrations
-* Quick access without leaving your current project context
+* Open your team from the agent sidebar.
+* The agents use your knowledge and your connected apps.
+* You stay in the current project.
**Option 2: Bulk Task Assignment**
-* Assign multiple tasks to your agent team simultaneously
-* Each agent handles tasks within their area of expertise
-* Automatic task distribution and progress tracking
-* Consolidated reporting on team performance
+* Assign many tasks to your team at the same time.
+* Each agent does the tasks in its area of expertise.
**Option 3: Conversational Collaboration**
-* Chat with your entire agent team as a unified group
-* Agents collaborate on complex problems and provide comprehensive solutions
-* Shared conversation history for context and continuity
-* Team members can participate alongside AI agents
+* Chat with your whole team in one chat.
+* The agents work together on hard problems.
+* The chat history stays available for context.
+* Your team members can join the chat with the agents.
#### Advanced Multi-Agent Features
**Memory and Context Management**
-* **Short-term memory** for immediate conversation context
-* **Long-term memory** for persistent knowledge and preferences
-* **Cross-agent knowledge sharing** for improved collaboration
-* **Conversation history** for iteration and refinement
+* **Short-term memory**: The context of the current chat
+* **Long-term memory**: The persistent knowledge of each agent
+* **Shared knowledge**: Agents can use the knowledge of the team
+* **Conversation history**: Earlier messages that you can use again
**Iterative Output Development**
-* Browse complete conversation histories
-* Refine and improve outputs through multiple iterations
-* Learn from previous interactions and feedback
-* Build upon successful approaches and templates
+* Browse the full history of a chat.
+* Improve an output over many replies.
+* Start again from approaches that worked.
**Team Communication and Sharing**
-* Shared agent chats for team-wide collaboration
-* Real-time collaboration on problem-solving
-* Collective brainstorming and idea generation
-* Distributed AI insights accessible to all team members
+* Share agent chats with your team.
+* Solve problems together in real time.
+* Brainstorm ideas as a group.
#### Managing Multi-Agent Teams
**Copy and Duplicate Teams**
-* Save successful team configurations as templates
-* Duplicate teams for similar projects or client work
-* Share team setups across workspaces and organizations
-* Version control for team configuration changes
+* Save a team setup that works as a template.
+* Duplicate a team for a similar project or client.
+* Share team setups with other workspaces.
**Sharing and Collaboration**
-* Share agent teams with team members and stakeholders
-* Granular permission controls for team access
-* Public sharing options for community contributions
-* Template marketplace for proven team configurations
+* Share agent teams with team members and stakeholders.
+* Control who can open each team.
+* Publish teams for the community.
#### Popular Multi-Agent Team Examples
@@ -519,81 +522,86 @@ Multi-agents are an evolution of Taskade's agentic workflow designed to:
| **Customer Success Team** | Support Agent + Documentation Specialist + Training Coordinator | Onboarding, support tickets, knowledge base |
| **Product Launch Team** | Market Researcher + Copywriter + Analytics Expert | Product announcements, launch campaigns, performance tracking |
| **Research & Development Team** | Data Analyst + Technical Writer + Innovation Specialist | Competitive analysis, technical documentation, trend reports |
-| **Operations Team** | Logistics Coordinator + Communication Manager + Reporting Analyst | Process optimization, team communication, performance dashboards |
+| **Operations Team** | Logistics Coordinator + Communication Manager + Reporting Analyst | Process improvement, team communication, performance dashboards |
### Human in the Loop Controls
-For sensitive operations, multi-agent teams incorporate human oversight:
+You can keep a human in control of sensitive work:
-* **Action Approval**: Agents request permission before critical actions
-* **Escalation Protocols**: Automatic human involvement for high-impact decisions
-* **Audit Trails**: Complete logging of all agent actions and approvals
-* **Override Capabilities**: Human ability to modify or stop agent processes
+* **Action Approval**: The agent asks for your approval before an action that you set to **Manual Approval**.
+* **Escalation**: You decide when a human must make the decision.
+* **Review**: You can read the actions of the agent in the chat.
+* **Stop**: You can reject an action before it runs.
## Agent Tools & Integrations
### Enable Agent Tools
-Agent Tools extend your multi-agent teams' capabilities by connecting to external services:
+Agent tools connect your agents and teams to other services. Turn them on in the **Tools** sub-tab of the agent's **Actions** tab:
-**Taskade Tools** - Create projects, update tasks, manage data **Web Search** - Find real-time information online **Email Integration** - Send emails through Gmail **Communication** - Post to Slack, Teams, Discord **Data Tools** - Work with Google Sheets, databases **Automation** - Trigger workflows and processes
+* **Taskade Tools**: Create projects, update tasks, and manage data.
+* **Web Search**: Find current information on the web.
+* **Email Integration**: Send emails with Gmail.
+* **Communication**: Post to Slack, Teams, and Discord.
+* **Data Tools**: Work with Google Sheets and databases.
+* **Automation**: Start automations and processes.
## Best Practices
### Writing Good Prompts
-* **Be specific:** "Write a welcome email for new customers" vs "Write an email"
-* **Give context:** Include relevant background information
-* **Set expectations:** Explain the tone, length, and format you want
-* **Provide examples:** Show the agent what good output looks like
+* **Be specific:** "Write a welcome email for new customers" is better than "Write an email".
+* **Give context:** Add the related background information.
+* **Set expectations:** Tell the agent the tone, length, and format.
+* **Give examples:** Show the agent a good output.
### Training Your Agent
-* **Start with quality content** - Upload your best documents first
-* **Update regularly** - Keep knowledge current and relevant
-* **Test thoroughly** - Try different questions to ensure good responses
-* **Iterate and improve** - Refine based on real usage
+* **Start with good content**: Upload your best documents first.
+* **Update regularly**: Keep the knowledge current.
+* **Test**: Ask different questions and read the answers.
+* **Improve**: Change the agent when real use shows a problem.
### Managing Multiple Agents
-* **Specialize agents** - Give each one a clear, focused role
-* **Use descriptive names** - Make it easy to pick the right agent
-* **Organize by function** - Group similar agents together
-* **Share successful agents** - Let teammates copy your best configurations
+* **Give each agent one role**: A clear role gives better answers.
+* **Use descriptive names**: Make it easy to pick the correct agent.
+* **Organize by function**: Keep similar agents together.
+* **Share agents that work**: Let teammates copy your best agents.
## Troubleshooting
### Agent Giving Wrong Answers?
-✅ Check if it has the right knowledge documents\
-✅ Make sure information is up-to-date\
-✅ Improve your questions - be more specific\
-✅ Add more training examples to the knowledge base
+✅ Make sure that the agent has the correct knowledge documents.\
+✅ Make sure that the information is current.\
+✅ Ask a more specific question.\
+✅ Add more examples to the knowledge.
### Agent Not Understanding Questions?
-✅ Rephrase your question more clearly\
-✅ Provide more context about what you need\
-✅ Check if the agent has relevant knowledge about the topic\
-✅ Try breaking complex questions into smaller parts
+✅ Write your question more clearly.\
+✅ Give more context about what you need.\
+✅ Make sure that the agent has knowledge about the topic.\
+✅ Divide a complex question into smaller questions.
### Agent Actions Not Working?
-✅ Verify tool permissions are enabled\
-✅ Check if external services (Gmail, Slack) are connected\
-✅ Make sure you approved any required actions\
-✅ Look for error messages in the agent chat
+✅ Make sure that the tool is turned on.\
+✅ Make sure that external services (Gmail, Slack) are connected.\
+✅ Make sure that you approved the actions that need approval.\
+✅ Look for error messages in the agent chat.
## Next Steps
-Ready to build your AI team?
+Do these steps to build your AI team:
-1. **Start with one agent** - Pick your biggest time drain
-2. **Train it well** - Upload quality knowledge and test thoroughly
-3. **Use it daily** - The more you use it, the more time you save
-4. **Expand gradually** - Add more agents as you see value
+1. **Start with one agent**: Pick the task that takes the most of your time.
+2. **Train it well**: Upload good knowledge and test the agent.
+3. **Use it daily**: Regular use saves more time.
+4. **Add more agents**: Add agents when you see value.
-**Need inspiration?**
+**Ideas and examples:**
* [Browse the AI Agent Gallery](https://taskade.com/agents) for community examples
* [Watch tutorial videos](https://youtube.com/taskade) for step-by-step guides
@@ -603,96 +611,96 @@ Ready to build your AI team?
### **Build Without Permission**
-You don't need permission to start. Not to ship. Not to define the system.
+You can start to build at once. You do not need approval to ship or to design your system.
-Your AI agents are just the beginning. Transform them into complete business applications with Genesis - the HyperCard moment for the AI era.
+Your AI agents can also become full business apps with Taskade Genesis. Taskade Genesis builds an app from one prompt.
[**Read the Manifesto →**](https://www.taskade.com/blog/build-without-permission) - Build Without Permission
### **Build Genesis Apps with Your Agents**
-Transform your AI agents into complete business applications. One prompt = one app. Your workspace = the backend. Your agents = the team.
+Turn your AI agents into business apps. One prompt makes one app. Your workspace holds the data, and your agents do the work.
* [**Genesis Getting Started →**](../genesis/getting-started.md) - Build your first app in 5 minutes
-* [**Genesis Examples →**](../genesis/examples-and-templates.md) - See what others have built
+* [**Genesis Examples →**](../genesis/examples-and-templates.md) - See apps that other people built
### **Connect Agents to Automations**
-Make your agents work automatically:
+Automations run your agents without a manual start:
-* [**Automation Getting Started →**](../automation/README.md) - Put agents on autopilot
+* [**Automation Getting Started →**](../automation/README.md) - Run agents on a schedule or an event
* [**Integration Guide →**](../automation/integrations.md) - Connect to 100+ services
#### **Build Agent-Centric Automations**
-Create automation workflows with your AI agents at the center:
+You can build automations that have an AI agent at the center.
**Access Agent Automation:**
-1. Navigate to the **Agents** tab in your workspace/folder
-2. Select an existing agent or create a new one
-3. Click **Edit Agent** in the top-right corner
-4. Go to the **Automation** tab on the left sidebar
+1. Open **AI Agents** in your workspace sidebar.
+2. Select an existing agent or create a new one.
+3. Open the agent's **Actions** tab.
+4. Open the **Automation** sub-tab.
**Create Agent Automation Flow:**
-1. **Choose a connector** from available automation triggers
-2. **Configure the trigger** to initiate your workflow
-3. **Set up your agent** as the central processing component
-4. **Add additional actions** before or after agent processing
-5. **Test and deploy** your agent automation
+1. **Choose a connector** from the available automation triggers.
+2. **Configure the trigger** that starts your automation.
+3. **Add your agent** as the main step.
+4. **Add more actions** before or after the agent step.
+5. **Test and turn on** your agent automation.
**Supported Triggers for Agent Automations:**
-* Task creation, updates, and completion events
-* Project and folder activity triggers
-* External service integrations (Gmail, Slack, etc.)
+* Task creation, updates, and completion
+* Activity in projects and folders
+* Events from external services (Gmail, Slack, and more)
* Time-based and scheduled triggers
* Custom webhook and API triggers
**Agent Automation Benefits:**
-* **Centralized Intelligence**: AI agents process and respond to complex scenarios
-* **Seamless Integration**: Connect AI capabilities directly to business workflows
-* **Autonomous Operation**: Agents handle sequences of tasks without manual intervention
-* **Contextual Processing**: Maintain conversation context and memory across interactions
-* **Scalable Automation**: Deploy intelligent workflows across multiple use cases
+* **Central knowledge**: The agent handles complex cases with its knowledge.
+* **Direct connection**: The AI step sits inside your business automations.
+* **No manual start**: The agent does a series of tasks without your input.
+* **Context**: The agent keeps the context and memory of the chat.
+* **Reuse**: You can use the same automation for many use cases.
### **Advanced Knowledge Management**
-Supercharge your agents with better knowledge:
+Give your agents better knowledge:
-* [**Knowledge Management Guide →**](../workspaces/knowledge-management.md) - Organize information effectively
-* [**Project Views →**](../workspaces/project-views.md) - Structure data for AI understanding
+* [**Knowledge Management Guide →**](../workspaces/knowledge-management.md) - Organize your information
+* [**Project Views →**](../workspaces/project-views.md) - Structure data so that the AI can read it
## Advanced Agent Features
### **Command Modes: How Your Agent Thinks**
-Command modes dictate how your AI agents approach tasks - whether by following specific instructions or by autonomously setting goals and working towards them.
+A command mode sets how an agent does a task. In one mode, the agent follows your instructions directly. In the other mode, the agent sets goals and makes a plan.
#### **Available Command Modes**
| Mode | Description | Best For | Example |
| --------------------- | ------------------------------------------- | ---------------------------------- | ----------------------------------------------------------- |
-| **Default** | Follows specific instructions directly | Simple, straightforward tasks | "Write an email to customer about order delay" |
-| **Plan and execute** | Sets goals, breaks the task into steps, and reasons through them | Complex multi-step goals | "Increase sales revenue" → Agent creates strategy plan |
+| **Default** | Follows specific instructions directly | Simple, direct tasks | "Write an email to customer about order delay" |
+| **Plan and execute** | Sets goals, divides the task into steps, and reasons through them | Complex goals with many steps | "Increase sales revenue" → Agent creates strategy plan |
#### **When to Use Plan & Execute Modes**
**✅ Use Plan & Execute for:**
-* Complex business problems without clear solutions
-* Goals that require research and analysis
-* Multi-step processes that need coordination
-* Strategic planning and decision-making
+* Complex business problems with no clear solution
+* Goals that need research and analysis
+* Processes with many steps that need coordination
+* Strategic planning and decisions
**❌ Use Default Mode for:**
-* Simple, direct tasks with clear instructions
-* Quick responses and straightforward actions
-* Tasks where you want specific output format
-* Routine operations with established procedures
+* Simple tasks with clear instructions
+* Quick replies and direct actions
+* Tasks where you want a specific output format
+* Routine work with known procedures
#### **Plan & Execute Examples**
@@ -731,126 +739,126 @@ Agent Process:
#### **Configuring Command Modes**
-1. **Edit your agent** and go to the **Commands** tab
-2. **Select a command** you want to modify
-3. **Choose the command mode** from the dropdown
-4. **Test the command** with sample input
-5. **Adjust as needed** based on results
+1. Open your agent's **Actions** tab, then the **Commands** sub-tab.
+2. Select the command that you want to change.
+3. Choose the command mode from the dropdown.
+4. Test the command with sample input.
+5. If the result is not correct, change the command and test again.
### **Bulk AI Commands: Process Multiple Tasks at Once**
-Apply AI commands to multiple tasks simultaneously for efficient batch processing.
+You can run one AI command on many tasks at the same time.
#### **How Bulk Commands Work**
-1. **Select multiple tasks** in your project (click one task, then hold Shift and click the last task)
-2. **Click AI Assistant** in the selection toolbar
-3. **Choose a bulk command** from the available options
-4. **AI reads the whole selection** and gives one result
-5. **Insert** the result after the tasks, or **Replace** the selected tasks
+1. Select the tasks in your project. Click the first task, then hold Shift and click the last task.
+2. Click **AI Assistant** in the selection toolbar.
+3. Choose a bulk command.
+4. Read the result. The AI reads the whole selection and gives one result.
+5. Click **Insert** to add the result after the tasks, or **Replace** to replace the selected tasks.
#### **Available Bulk AI Commands**
| Command | Action | Best For | Example |
| ------------------------------- | ------------------------------------- | -------------------------------------- | -------------------------------------------- |
-| **🟪 Research** | Search the web to learn about topics | Market research, fact-checking | Research competitors for each product |
-| **🟪 SEO** | Fetch top Google Search results | Content optimization, keyword research | Find SEO keywords for blog topics |
-| **🟠 Summarize** | Shrink down long discussions or notes | Meeting notes, document reviews | Summarize customer feedback responses |
-| **🟤 Turn into Action Items** | Convert ideas into actionable steps | Project planning, brainstorming | Turn feature requests into development tasks |
-| **🟡 Prioritize** | Prioritize tasks based on importance | Task management, sprint planning | Rank support tickets by urgency |
-| **🔴 Continue Writing** | Expand on existing content | Content creation, documentation | Expand product descriptions |
-| **⚪️ Brainstorm** | Develop ideas in seconds | Creative projects, problem-solving | Generate marketing campaign ideas |
-| **🔵 Rewrite** | Refine words for clarity and impact | Content editing, communication | Improve email templates |
-| **🟢 Outline** | Turn ideas into structured plans | Project planning, content strategy | Create article outlines from topics |
-| **🟥 Fix Spelling and Grammar** | Edit out mistakes | Content review, documentation | Clean up imported content |
+| **🟪 Research** | Search the web about a topic | Market research, fact checks | Research competitors for each product |
+| **🟪 SEO** | Get the top Google Search results | Content improvement, keyword research | Find SEO keywords for blog topics |
+| **🟠 Summarize** | Make long discussions or notes short | Meeting notes, document reviews | Summarize customer feedback responses |
+| **🟤 Turn into Action Items** | Change ideas into steps | Project planning, brainstorms | Turn feature requests into development tasks |
+| **🟡 Prioritize** | Sort tasks by importance | Task management, sprint planning | Rank support tickets by urgency |
+| **🔴 Continue Writing** | Add to existing content | Content creation, documentation | Expand product descriptions |
+| **⚪️ Brainstorm** | Find new ideas fast | Creative projects, problem solving | Generate marketing campaign ideas |
+| **🔵 Rewrite** | Make text clearer | Content editing, communication | Improve email templates |
+| **🟢 Outline** | Change ideas into structured plans | Project planning, content strategy | Create article outlines from topics |
+| **🟥 Fix Spelling and Grammar** | Correct mistakes | Content review, documentation | Clean up imported content |
#### **Custom Agent Bulk Commands**
-You can also use your custom AI agent commands with multiple tasks:
+You can also run your custom agent commands on many tasks:
-1. **Create custom commands** in your agent settings
-2. **Select multiple tasks** that need the same processing
-3. **Choose your custom agent command** from the bulk options
-4. **AI applies your specialized command** to all selected tasks
+1. Create custom commands in the agent's **Actions** tab.
+2. Select the tasks that need the same command.
+3. Choose your custom agent command from the bulk options.
+4. Read the result. The AI applies your command to all selected tasks.
**Example Use Cases:**
-* **Sales Agent**: Qualify multiple leads simultaneously
-* **Content Agent**: Generate social media posts for multiple products
-* **Support Agent**: Categorize multiple customer inquiries
-* **Research Agent**: Analyze multiple competitor websites
+* **Sales Agent**: Qualify many leads at the same time.
+* **Content Agent**: Write social media posts for many products.
+* **Support Agent**: Sort many customer questions.
+* **Research Agent**: Analyze many competitor websites.
#### **Bulk Command Best Practices**
**✅ Do:**
-* Group similar tasks for bulk processing
-* Review results and refine as needed
-* Use consistent formatting across tasks
-* Test commands on small batches first
+* Group similar tasks in one bulk command.
+* Review the results and improve them.
+* Use the same format in all tasks.
+* Test a command on a small batch first.
-**❌ Don't:**
+**❌ Do not:**
-* Mix different task types in one bulk operation
-* Apply complex commands to too many tasks at once
-* Skip reviewing AI-generated results
-* Use bulk commands for tasks requiring individual attention
+* Mix different types of tasks in one bulk command.
+* Run a complex command on too many tasks at the same time.
+* Accept AI results without a review.
+* Use bulk commands for tasks that need individual attention.
***
## **Mobile Agent Development**
-Build and train AI agents directly from your mobile device, making agent development truly accessible anywhere.
+You can build and train AI agents on your mobile device.
### **Mobile Agent Creation**
| **Feature** | **Capability** | **Use Case** |
| --------------------------- | ---------------------------------- | --------------------------------- |
-| **Quick Agent Setup** | Create agents with voice prompts | Build agents while commuting |
+| **Quick Agent Setup** | Create agents with voice prompts | Build agents while you travel |
| **Mobile Knowledge Upload** | Add photos, documents, voice memos | Train agents with real-world data |
-| **Agent Testing** | Test agent responses on mobile | Validate agent behavior instantly |
-| **Mobile Commands** | Create custom commands via touch | Build agent tools on the go |
+| **Agent Testing** | Test agent responses on mobile | See agent behavior at once |
+| **Mobile Commands** | Create custom commands with touch | Build agent commands anywhere |
### **Mobile-Specific Agent Tools**
| **Tool** | **Mobile Advantage** | **Business Impact** |
| ---------------------- | ----------------------------- | ----------------------------- |
-| **Photo Analysis** | Analyze images from camera | Instant product cataloging |
-| **Voice Commands** | Speak to agents naturally | Hands-free agent interaction |
+| **Photo Analysis** | Analyze images from camera | Fast product catalogs |
+| **Voice Commands** | Speak to agents naturally | Hands-free agent chats |
| **Push Notifications** | Receive agent alerts anywhere | Real-time business monitoring |
### **Mobile Agent Use Cases**
**Field Service Agent:**
-* Take photos of equipment issues
-* Get instant diagnostic suggestions
-* Create work orders automatically
-* Access repair manuals contextually
+* Take photos of equipment problems.
+* Get diagnostic suggestions at once.
+* Create work orders.
+* Open the related repair manuals.
**Sales Agent:**
-* Capture client meeting notes via voice
-* Get real-time product recommendations
-* Access pricing and availability instantly
-* Generate follow-up tasks automatically
+* Record client meeting notes with your voice.
+* Get product recommendations in real time.
+* Get pricing and availability at once.
+* Create follow-up tasks.
## **AI Team Generator: Multi-Agent Teams**
-Create entire teams of specialized AI agents with a single prompt using the revolutionary AI Team Generator.
+The AI Team Generator creates a team of specialized AI agents from one prompt.
### **What is AI Team Generator?**
-The AI Team Generator lets you create complete multi-agent teams instantly. Just describe your use case, and Taskade automatically generates a fully configured AI Team with tailored roles, behaviors, and specialized toolkits.
+You describe your use case, and Taskade generates an AI Team. Each agent in the team gets its own role, instructions, and tools.
### **When to Use AI Team Generator**
-| **Scenario** | **Why It's Perfect** |
+| **Scenario** | **Why It Helps** |
| -------------------------- | ---------------------------------------------------- |
-| **Complex Workflows** | Multiple agents tackle different parts of a workflow |
-| **Uncertain Structure** | AI designs the optimal team composition for you |
-| **Time Constraints** | Save hours of manual agent configuration |
-| **Best Practice Learning** | See how experts structure multi-agent systems |
+| **Complex Workflows** | Many agents do different parts of the work |
+| **Unknown Structure** | Taskade designs the team for you |
+| **Little Time** | You do not configure each agent by hand |
+| **Learning** | You see how a multi-agent team is structured |
### **Example AI Team Generations**
@@ -858,29 +866,31 @@ The AI Team Generator lets you create complete multi-agent teams instantly. Just
**Generated Team:**
-* **🎯 Product Manager Agent** - Roadmap planning, feature prioritization
-* **📊 Market Research Agent** - Competitor analysis, user research
-* **✍️ Content Marketing Agent** - Blog posts, social media, email campaigns
-* **📈 Growth Marketing Agent** - SEO optimization, conversion tracking
-* **🎨 Design Agent** - Visual assets, UI/UX feedback
-* **💬 Customer Success Agent** - Onboarding flows, support documentation
+* **🎯 Product Manager Agent**: Roadmap planning, feature priorities
+* **📊 Market Research Agent**: Competitor analysis, user research
+* **✍️ Content Marketing Agent**: Blog posts, social media, email campaigns
+* **📈 Growth Marketing Agent**: SEO, conversion tracking
+* **🎨 Design Agent**: Visual assets, UI/UX feedback
+* **💬 Customer Success Agent**: Onboarding flows, support documentation
**Prompt:** _"Build a legal compliance team for healthcare"_
**Generated Team:**
-* **⚖️ Compliance Officer Agent** - HIPAA regulations, audit preparation
-* **📋 Documentation Agent** - Policy creation, procedure updates
-* **🔍 Risk Assessment Agent** - Vulnerability analysis, mitigation strategies
-* **📊 Reporting Agent** - Compliance reports, regulatory filings
-* **🎓 Training Agent** - Staff education, certification tracking
+* **⚖️ Compliance Officer Agent**: HIPAA regulations, audit preparation
+* **📋 Documentation Agent**: Policy creation, procedure updates
+* **🔍 Risk Assessment Agent**: Vulnerability analysis, mitigation strategies
+* **📊 Reporting Agent**: Compliance reports, regulatory filings
+* **🎓 Training Agent**: Staff education, certification tracking
### **How It Works**
-1. **Describe Your Need** - "I need a team for \[specific use case]"
-2. **AI Analyzes Requirements** - Identifies roles, skills, and workflows needed
-3. **Generates Specialized Agents** - Each with custom instructions and tools
-4. **Ready to Deploy** - Start using your team immediately
+The generator does these steps:
+
+1. **You describe your need**: For example, "I need a team for \[specific use case]".
+2. **Taskade analyzes the request**: It finds the roles, skills, and work that the team needs.
+3. **Taskade generates the agents**: Each agent gets custom instructions and tools.
+4. **The team is ready**: You can use the team at once.
### **How to Use AI Team Generator**
@@ -890,14 +900,14 @@ The AI Team Generator lets you create complete multi-agent teams instantly. Just
#### **Access the Generator**
-1. Open **AI Teams** in your workspace sidebar
-2. Click **Generate**
-3. Describe the work in the **Generate Your AI Agent Team** dialog
+1. Open **AI Teams** in your workspace sidebar.
+2. Click **Generate**.
+3. Describe the work in the **Generate Your AI Agent Team** dialog.
#### **Craft Your Prompt**
{% hint style="info" %}
-The quality of your prompt determines the quality of your generated team. Be specific about your use case, industry, and requirements.
+A better prompt gives a better team. A good prompt names your use case, your industry, and your requirements.
{% endhint %}
**Effective Prompt Examples:**
@@ -909,18 +919,18 @@ The quality of your prompt determines the quality of your generated team. Be spe
### **Editing and Customizing Generated Teams**
{% hint style="warning" %}
-All generated teams can be fine-tuned after creation to better fit your specific needs.
+You can change every generated team after Taskade creates it.
{% endhint %}
#### **Available Customizations**
-| **Customization Type** | **What You Can Modify** | **Impact** |
+| **Customization Type** | **What You Can Change** | **Result** |
| ------------------------ | ------------------------------------------------ | ---------------------------------------- |
-| **Agent Roles** | Change titles, responsibilities, and specialties | Better alignment with your workflow |
-| **Instructions** | Adjust behavior, tone, and response guidelines | More accurate and helpful responses |
-| **Tools & Integrations** | Add/remove connected services and APIs | Enhanced capabilities and automation |
-| **Knowledge Base** | Upload additional training materials | More informed and knowledgeable agents |
-| **Communication Style** | Modify how agents interact and collaborate | Better team dynamics and user experience |
+| **Agent Roles** | Titles, responsibilities, and specialties | A better fit with your work |
+| **Instructions** | Behavior, tone, and reply guidelines | More correct and useful replies |
+| **Tools & Integrations** | Add or remove connected services and APIs | More actions and automation |
+| **Knowledge Base** | Upload more training materials | Agents that know more |
+| **Communication Style** | How agents talk and work together | Better teamwork and user experience |
#### **Advanced Editing Features**
@@ -928,25 +938,25 @@ All generated teams can be fine-tuned after creation to better fit your specific
{% step %}
#### Access Team Settings
-Click **···** next to your team, then **Edit team**
+Click **···** next to your team, then click **Edit team**.
{% endstep %}
{% step %}
#### Modify Agent Configuration
-Select individual agents to adjust their parameters
+Select an agent and change its settings.
{% endstep %}
{% step %}
#### Test Changes
-Use the **preview chat** to test modified behavior
+Use the **preview chat** to test the changed behavior.
{% endstep %}
{% step %}
#### Save and Deploy
-Apply changes and make the team available for use
+Save the changes. The team is then ready to use.
{% endstep %}
{% endstepper %}
@@ -958,77 +968,74 @@ Apply changes and make the team available for use
#### **Method 1: Direct Team Chat**
-* Access through the **AI Teams** tab
-* Chat with the entire team as a unified group
-* Pick **Auto**, **Everyone**, **Manual**, or **Orchestrate** to decide which agents reply
-* Perfect for brainstorming and complex problem-solving
+* Open the team in **AI Teams**.
+* Chat with the whole team in one chat.
+* Pick **Auto**, **Everyone**, **Manual**, or **Orchestrate** to decide which agents reply.
+* Use the team chat for brainstorms and hard problems.
#### **Method 2: Project Integration**
-* Assign teams to specific projects
-* Agents work within project context
-* Automatic task generation and updates
-* Seamless integration with existing workflows
+* Open the team in the chat of a project.
+* The agents use the project as context.
#### **Method 3: Task Assignment**
-* Type **/** and the team name in a task to assign it
-* Taskade picks the single best agent from the team for each task
-* Insert or replace the task with the result
+* In a task, type **/** and the team name to assign the task.
+* Taskade picks the single best agent from the team for each task.
+* Insert the result after the task, or replace the task with the result.
#### **Method 4: Automation Integration**
-* Connect teams to automation workflows
-* Trigger team actions based on events
-* Automated responses and follow-ups
-* Continuous background processing
+* Add the **Ask an agent team** step to an automation.
+* Start the team from an event, such as a new form entry.
+* Send the replies of the team to the next steps.
### **AI Team Use Cases and Examples**
| **Industry** | **Team Type** | **Key Agents** | **Benefits** |
| --------------- | ------------------ | ------------------------------------- | ------------------------------ |
-| **Marketing** | Content Creation | Research, Writing, Design, SEO | Complete content pipeline |
-| **Sales** | Lead Management | Qualification, Follow-up, Analytics | Streamlined sales process |
-| **Support** | Customer Service | Triage, Resolution, Feedback | 24/7 customer support |
-| **Development** | Project Management | Planning, Coding, Testing, Deployment | Accelerated development cycles |
-| **HR** | Recruitment | Sourcing, Screening, Interviewing | Efficient hiring process |
+| **Marketing** | Content Creation | Research, Writing, Design, SEO | Full content process |
+| **Sales** | Lead Management | Qualification, Follow-up, Analytics | Faster sales process |
+| **Support** | Customer Service | Triage, Resolution, Feedback | Customer support at all hours |
+| **Development** | Project Management | Planning, Coding, Testing, Deployment | Faster development cycles |
+| **HR** | Recruitment | Sourcing, Screening, Interviewing | Faster hiring |
### **Best Practices for AI Teams**
{% hint style="info" %}
-**Optimization Strategies for Maximum Impact**
+**Tips for Better Teams**
{% endhint %}
#### **Team Composition**
-* **Start Small**: Begin with 3-5 agents for focused teams
-* **Role Clarity**: Ensure each agent has distinct, non-overlapping responsibilities
-* **Skill Balance**: Include a mix of specialized and general-purpose agents
+* **Start Small**: Start with 3 to 5 agents in one team.
+* **Clear Roles**: Give each agent different responsibilities.
+* **Skill Balance**: Use a mix of specialized and general agents.
#### **Performance Tuning**
-* **Prompt Refinement**: Iterate on team prompts for better results
-* **Tool Integration**: Connect relevant external services and APIs
-* **Knowledge Updates**: Regularly refresh agent knowledge bases
+* **Prompt Refinement**: Improve your team prompts over time.
+* **Tool Integration**: Connect the external services and APIs that the team needs.
+* **Knowledge Updates**: Update the knowledge of each agent at regular intervals.
#### **Workflow Integration**
-* **Clear Triggers**: Define when and how teams should be activated
-* **Output Standards**: Establish consistent formatting and quality expectations
-* **Feedback Loops**: Implement review processes for continuous improvement
+* **Clear Triggers**: Decide when and how the team starts.
+* **Output Standards**: Set the format and quality that you expect.
+* **Feedback Loops**: Review the output of the team at regular intervals.
### **Troubleshooting AI Teams**
| **Issue** | **Common Causes** | **Solutions** |
| --------------------------- | ------------------------- | ---------------------------------------------- |
-| **Uncoordinated Responses** | Overlapping agent roles | Clarify responsibilities in team configuration |
-| **Slow Response Times** | Too many agents active | Reduce team size or optimize prompts |
-| **Inconsistent Quality** | Unclear instructions | Refine agent prompts and guidelines |
-| **Integration Problems** | API/authentication issues | Verify tool connections and permissions |
+| **Uncoordinated Responses** | Agent roles overlap | Give each agent a clear role in the team setup |
+| **Slow Response Times** | Too many agents reply | Use fewer agents or shorter prompts |
+| **Inconsistent Quality** | Unclear instructions | Improve agent prompts and guidelines |
+| **Integration Problems** | API or sign-in problems | Make sure that tool connections and permissions are correct |
***
-**💡 Pro Tip:** Your best agent team is one that solves a real problem you face daily. Start with a specific use case, measure results, and expand as you discover new possibilities!
+**💡 Pro Tip:** The best agent team solves a real problem that you have each day. Start with one use case, measure the results, and then add more.
---
diff --git a/genesis-living-system-builder/ai-features/multi-agents.md b/genesis-living-system-builder/ai-features/multi-agents.md
index da91b01..ec9deef 100644
--- a/genesis-living-system-builder/ai-features/multi-agents.md
+++ b/genesis-living-system-builder/ai-features/multi-agents.md
@@ -9,9 +9,9 @@ description: >-
## Overview
-Multi-Agents lets you **assign multiple AI agents to work together** on projects, tasks, and conversations. Instead of relying on a single generalist AI, you create teams of specialized agents that collaborate — each bringing domain expertise, unique knowledge, and specific tools.
+With multi-agents, **several AI agents work together** on projects, tasks, and chats. Each agent is a specialist with its own knowledge and tools. Together they do more than one general agent can do.
-Specialized agents, Taskade EVE, and people coordinate around one shared project.
+Specialized agents, Taskade EVE, and people work together around one shared project.
```mermaid
graph TD
@@ -22,7 +22,7 @@ graph TD
```
-> **Workspace DNA impact:** Multi-agent teams amplify the Intelligence pillar. Each agent contributes +10 points to your [Intelligence Score](../genesis/how-genesis-works.md#workspace-intelligence-score), and teams create compounding intelligence through specialization.
+> **Workspace DNA impact:** Multi-agent teams make the Intelligence pillar stronger. Each agent adds +10 points to your [Intelligence Score](../genesis/how-genesis-works.md#workspace-intelligence-score). A team of specialists gives better results than one general agent.
>
> **Plan note:** Multi-Agent Teams are available on Pro plans and up.
@@ -32,36 +32,36 @@ graph TD
### Method 1: Build from Scratch
-1. Open the **Agents** tab in your workspace
-2. Click **Create Agent**
-3. Configure: name, persona, tone, knowledge, tools
-4. Repeat for each specialized agent in your team
+1. Open **AI Agents** in your workspace sidebar.
+2. Click **Create** to start a blank agent. You can also click **Generate** to describe the agent, or **Templates** to pick one.
+3. Add the name, persona, tone, knowledge, and tools.
+4. Do these steps again for each specialist in your team.
### Method 2: Use Templates
-1. Browse the agent template library
-2. Select pre-built agent templates for each role
-3. Customize each agent for your specific use case
+1. Open the agent template library.
+2. Select a template for each role.
+3. Change each agent for your use case.
-### Method 3: Auto-Generate a Team with AI
+### Method 3: Generate a Team with AI
-The fastest way to create a multi-agent team:
+This is the fastest way to make a multi-agent team:
-1. Open the **AI Teams** tab
-2. Click **Generate**
-3. Describe your team in plain English (e.g., "I need a marketing team with a content writer, SEO specialist, social media manager, and analytics analyst")
-4. AI generates the complete team with specialized agents, knowledge sources, and tool configurations
-5. Review and customize each agent as needed
+1. In your workspace sidebar, open **AI Teams**.
+2. Click **Generate**. You can also click **+** (**Create AI team**) next to **AI Teams** and pick **Generate**.
+3. Describe your team in plain English (for example, "I need a marketing team with a content writer, SEO specialist, social media manager, and analytics analyst").
+4. Press **Enter**. Taskade creates the specialized agents and gives each one a role and tools.
+5. Review each agent, and change it if necessary.
-> **Pro tip:** Auto-generated teams include appropriate personas, knowledge connections, and tool assignments based on your description. You can always fine-tune each agent individually afterward.
+> **Tip:** A generated team gets personas, knowledge connections, and tools that match your description. You can change each agent later.
**Example team setup:**
| Agent | Specialty | Knowledge | Tools |
|---|---|---|---|
-| **Research Agent** | Data gathering, web search | Industry reports, competitor URLs | Search Web, Scrape Webpage |
+| **Research Agent** | Data collection, web search | Industry reports, competitor URLs | Search the web, Read a web page |
| **Analyst Agent** | Data analysis, insights | Historical data projects, KPIs | Google Sheets, Custom Fields |
-| **Writer Agent** | Content creation, copywriting | Brand guidelines, tone docs | Google Docs, WordPress |
+| **Writer Agent** | Content, copywriting | Brand guidelines, tone documents | Google Docs, WordPress |
| **Coordinator Agent** | Task management, routing | Team structure, processes | Slack, Google Calendar, Taskade Actions |
---
@@ -70,23 +70,26 @@ The fastest way to create a multi-agent team:
### Assign a Single Agent
-1. Open a project or task
-2. Click the **Agent** icon
-3. Select an agent from your workspace
-4. The agent now has context for that specific project/task
+1. Open a project or a task.
+2. Click the **Agent** icon.
+3. Select an agent from your workspace.
+
+The agent now has the context of that project or task.
### Assign Multiple Agents
-1. Open a project
-2. Click the **Agent** icon
-3. Select multiple agents for the project
-4. Each agent applies its specialty to the project
+1. Open a project.
+2. Click the **Agent** icon.
+3. Select more than one agent for the project.
+
+Each agent uses its specialty on the project.
### Bulk-Assign Agents
-1. Select multiple tasks or projects
-2. Use bulk actions to assign agents
-3. Agents are assigned to all selected items
+1. Select many tasks or projects.
+2. Use the bulk actions to assign agents.
+
+Taskade assigns the agents to all selected items.
---
@@ -95,10 +98,10 @@ The fastest way to create a multi-agent team:
| Feature | Detail |
|---|---|
| **Access** | Click the Agent icon in the sidebar |
-| **Switch agents** | Toggle between assigned agents mid-conversation |
-| **Context awareness** | Agent sees the project/task you're currently viewing |
-| **Tool access** | All enabled tools are available in the sidebar |
-| **History** | Conversation history preserved per agent |
+| **Switch agents** | Change between assigned agents in the same chat |
+| **Context** | The agent sees the project or task that you have open |
+| **Tool access** | All enabled tools work in the sidebar |
+| **History** | Taskade keeps the chat history of each agent |
@@ -106,28 +109,28 @@ The fastest way to create a multi-agent team:
## Iterate on Output
-Multi-agent output can be refined iteratively:
+You can improve the output of multi-agents step by step:
| Technique | Example |
|---|---|
-| **Sequential refinement** | Research Agent drafts → Writer Agent refines → Reviewer Agent checks |
-| **Split generation** | Multiple agents each take a different section, one after another |
-| **Feedback loop** | Ask one agent to critique another agent's output |
-| **Role switching** | "Now have the Analyst Agent review what the Writer Agent produced" |
+| **Step-by-step improvement** | Research Agent drafts → Writer Agent improves → Reviewer Agent reviews |
+| **Split work** | Each agent writes a different section, one after another |
+| **Review loop** | Ask one agent to review the output of another agent |
+| **Role change** | "Now have the Analyst Agent review what the Writer Agent produced" |
---
## Chat with Agents & Your Team
-Multi-agent chats support both AI agents and human team members in the same conversation:
+AI agents and team members can be in the same chat:
| Participant | Role |
|---|---|
-| **AI Agents** | Respond based on knowledge and tools |
-| **Human team members** | Provide direction, make decisions, add context |
-| **EVE** | Orchestrates and can invoke other agents |
+| **AI Agents** | Reply from their knowledge and tools |
+| **Human team members** | Give direction, make decisions, add context |
+| **Taskade EVE** | Orchestrates the chat and can call other agents |
-> **Tip:** @mention specific agents in chat to direct questions or tasks to them.
+> **Tip:** Type @ and the name of an agent to send a question or a task to that agent.
---
@@ -135,9 +138,9 @@ Multi-agent chats support both AI agents and human team members in the same conv
| Action | How | Result |
|---|---|---|
-| **Copy an agent** | Agent settings → Duplicate | Creates an identical copy you can customize |
-| **Copy a team** | Duplicate the project with agent assignments | All agents and their assignments are copied |
-| **Share across workspaces** | Publish agent publicly → Clone in another workspace | Agent config is replicated (knowledge may need re-linking) |
+| **Copy an agent** | Agent settings → Duplicate | Makes a copy that you can change |
+| **Copy a team** | Duplicate the project with its agent assignments | Taskade copies all agents and their assignments |
+| **Share across workspaces** | Publish the agent → Clone it in another workspace | Taskade copies the agent settings. If the knowledge does not connect, connect it again |
---
@@ -145,10 +148,10 @@ Multi-agent chats support both AI agents and human team members in the same conv
| Sharing Method | Scope | How |
|---|---|---|
-| **Within workspace** | Team members | Agents are available to all workspace members by default |
-| **Cross-workspace** | Other workspaces | Publish publicly → Clone |
-| **Public** | Anyone | Publish with public toggle |
-| **Embedded in Taskade Genesis app** | App users | Add agent chat widget to your Genesis app |
+| **Within workspace** | Team members | All workspace members can use agents by default |
+| **Cross-workspace** | Other workspaces | Publish the agent → Clone it |
+| **Public** | Anyone | Turn on **Publish Agent** |
+| **Embedded in Taskade Genesis app** | App users | Add an agent chat widget to your Taskade Genesis app |
---
@@ -156,23 +159,23 @@ Multi-agent chats support both AI agents and human team members in the same conv
| Pattern | Description | Best For | Example |
|---|---|---|---|
-| **Sequential Pipeline** | Agent A → Agent B → Agent C (output flows linearly) | Content creation, data processing | Research → Write → Review |
-| **Hub and Spoke** | Central coordinator routes to specialists | Customer support, complex queries | Coordinator → routes to Sales/Support/Billing |
-| **Fan-Out** | Multiple agents each answer the same request (in an AI Team, they reply one after another) | Report generation, data analysis | Finance + Marketing + Operations agents each analyze their domain |
-| **Debate / Review** | Agents critique each other's work | Quality control, decision-making | Proposal Agent creates → Critic Agent reviews → Proposal Agent revises |
-| **Escalation Chain** | Start with generalist, escalate to specialist | Support tiers, triage | Tier 1 Agent → Tier 2 Specialist → Human escalation |
+| **Sequential Pipeline** | Agent A → Agent B → Agent C (the output moves in a line) | Content, data processing | Research → Write → Review |
+| **Hub and Spoke** | A central coordinator sends work to specialists | Customer support, difficult questions | Coordinator → sends to Sales, Support, or Billing |
+| **Fan-Out** | Many agents answer the same request (in an AI Team, they reply one after another) | Reports, data analysis | Finance, Marketing, and Operations agents each analyze their area |
+| **Debate / Review** | Agents review the work of other agents | Quality control, decisions | Proposal Agent writes → Critic Agent reviews → Proposal Agent revises |
+| **Escalation Chain** | A general agent starts, and a specialist takes over | Support levels, triage | Level 1 Agent → Level 2 Specialist → Human |
---
## What's Next
-| Guide | What You'll Learn |
+| Guide | What you learn |
|---|---|
-| [Custom AI Agents](./ai-agents-getting-started.md) | Build individual agents |
-| [Agent Knowledge & Memory](./agent-knowledge.md) | Train agents with domain knowledge |
-| [Tools for AI Agents](./agent-tools.md) | Add capabilities to each agent |
-| [Automations](../automation/README.md) | Connect agents to automated workflows |
-| [Autonomous Agents (API)](../../apis-living-system-development/autonomous-agents.md) | Orchestrate multi-agent systems programmatically |
+| [Custom AI Agents](./ai-agents-getting-started.md) | Build single agents |
+| [Agent Knowledge & Memory](./agent-knowledge.md) | Give agents knowledge for their area |
+| [Tools for AI Agents](./agent-tools.md) | Add tools to each agent |
+| [Automations](../automation/README.md) | Connect agents to automations |
+| [Autonomous Agents (API)](../../apis-living-system-development/autonomous-agents.md) | Orchestrate multi-agent systems from code |
---
diff --git a/genesis-living-system-builder/automation/README.md b/genesis-living-system-builder/automation/README.md
index 2dedee1..a93d1a4 100644
--- a/genesis-living-system-builder/automation/README.md
+++ b/genesis-living-system-builder/automation/README.md
@@ -1,11 +1,11 @@
---
description: >-
- Automate work 24/7 with Taskade's trigger-and-action engine — connect 100+ integrations and let AI agents run your workflows from start to finish.
+ Build Taskade automations that run 24/7. A trigger starts each one, and its steps use AI agents and 100+ integrations to do the work.
---
# Automations Overview
-Automations let you build trigger-and-action workflows that run on their own — connecting Taskade to your tools and letting agents act 24/7.
+Automations are workflows that run on their own. A trigger starts each automation, and its steps do the work in Taskade and in your connected tools.
## I want to…
@@ -18,17 +18,17 @@ Automations let you build trigger-and-action workflows that run on their own —
| See every available connector | [Integration Options](integrations.md) |
| Start from a ready-made recipe | [Automation Recipes](recipes.md) |
-**Act. Execute. Evolve.** Taskade Automations are the Execution layer of Workspace DNA. They turn static workflows into systems that react to events and act for you.
+Automations are the Execution layer of Workspace DNA. They react to events and do the work for you.
{% hint style="success" %}
-**Execution Layer** — Automations act on events in your workspace and your connected tools. Every run is logged, so you can review it and refine the flow.
+**Execution Layer**: Automations act on events in your workspace and in your connected tools. Taskade keeps a log of every run. You can examine each run and improve the flow.
{% endhint %}
## What are automations?
-Automations are the Execution layer of your workspace. They are how your workspace can **react**, **decide**, and **execute** when something happens.
+An automation lets your workspace react to an event, make a decision, and do the work.
-Every automation follows the same shape: a trigger starts it, an optional condition routes it, and actions do the work.
+Every automation has the same shape. A trigger starts it. An optional condition sends it on the correct path. Then the actions do the work.
```mermaid
flowchart TD
@@ -39,72 +39,76 @@ flowchart TD
```
-Canonical foundation: [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
+For the full model, read [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
-### What you use it for
+### What you can do with automations
-* **Notifications**: send Slack/Teams/email alerts.
-* **Routing**: assign work based on content, priority, or owner.
-* **Sync**: keep CRMs, Sheets, and databases updated.
-* **Reporting**: scheduled summaries and digests.
-* **Approvals**: add a human-in-the-loop before publishing or sending.
+* **Notifications**: Send alerts to Slack, Microsoft Teams, or email.
+* **Routing**: Assign work by content, priority, or owner.
+* **Sync**: Keep CRMs, spreadsheets, and databases up to date.
+* **Reports**: Send scheduled summaries and digests.
+* **Approvals**: Add a **Wait for approval** step. The run waits until a teammate approves or rejects it.
### Core building blocks
#### Triggers
-Triggers start workflows. Common ones are schedules, forms, webhooks, and project events.
+A trigger starts a workflow. Common triggers are schedules, forms, webhooks, and project events.
#### Actions
-Actions do work. Create tasks, update fields, call webhooks, send notifications, and more.
+An action does the work. Actions create tasks, update fields, call webhooks, and send notifications.
-#### Conditions and branching
+#### Conditions and branches
-Use branches to route different cases. Example: urgent → page someone, normal → add to queue.
+A **Branch** step sends different cases on different paths. For example, an urgent case alerts a person, and a normal case goes into a queue.
#### AI steps
-Use an agent to classify, extract, or draft. Then route or act on the structured result.
+An **Ask an agent** step can sort, extract, or write text. To get typed fields back, set **Output Type** to **Structured Output**. Later steps can then use each field.
+
+#### Failed steps
+
+If an app is busy or down for a few minutes, check **Retry if this step fails** on that step. To let the run go on after a failure, check **Continue if this step fails**.
### Getting started (fast)
-1. Pick one workflow you want end-to-end.
-2. Choose one trigger.
-3. Add 1–3 actions.
-4. Test with real data.
-5. Add branching and AI only if needed.
+1. Select one workflow that you want to automate from start to end.
+2. Select one trigger.
+3. Add 1 to 3 actions.
+4. Test the automation with real data.
+5. Add branches and AI steps only when you need them.
### Typical workflows
-* Form submission → classify → create task → notify channel
-* Webhook → normalize data → update CRM → log result
-* Schedule → build weekly summary → send digest
+* Form submission → sort with AI → create task → send a channel message
+* Webhook → clean the data → update the CRM → log the result
+* Schedule → write a weekly summary → send a digest
### Available Triggers
**Project Events:**
-* **Task Added** — when a new task is created
-* **Task Completed** — when a task is marked done
-* **Task Assigned** — when a task is assigned to someone
-* **Task Due** — when a task reaches its due date
-* **New Comment** — when a comment is added
-* **New Due Date** — when a due date is set
-* **Custom Field Updated** — when a custom field value changes
-* **Project Completed** — when all tasks are done
-* **Public Agent Chat Ended** — when a public agent conversation finishes
-* **Task Moved** — when a task moves to another block within a project
-* **Project Created** — when a new project is created in a space
-* **Due Date Removed** — when a due date is cleared from a task
-* **File Added to Media** — when a file is uploaded to the media library
+* **Task Added**: a person or a step adds a task
+* **Task Completed**: a task is marked as done
+* **Task Assigned**: a person gets a task
+* **Task Due**: a task reaches its due date
+* **New Comment**: a person adds a comment
+* **New Due Date**: a task gets a due date
+* **Custom Field Updated**: the value of a custom field changes
+* **Project Completed**: all tasks in the project are done
+* **Public Agent Chat Ended**: a public agent conversation ends
+* **Task Moved**: a task moves to another block in the same project
+* **Project Created**: a new project starts in a space
+* **Due Date Removed**: a task loses its due date
+* **File Added to Media**: a person uploads a file to Media
**External Triggers:**
-* **Webhook** — receive data from any external service via HTTP POST (Pro and up)
-* **Form** — custom input forms with text, email, date, and time fields
-* **Schedule** — run every X minutes (sub-hourly intervals, paid plans), hourly, daily, weekly, or monthly
-* **Mailhook** — trigger from incoming emails
+* **Webhook**: receives data from an outside service with an HTTP POST request (Pro and up)
+* **Form**: a person submits a form that you build
+* **Schedule**: runs every few minutes (paid plans), every hour, day, week, or month
+* **Mailhook**: an email arrives at the address of the automation
**Integration Triggers** (from connected services):
@@ -112,7 +116,7 @@ Calendly, Slack, Typeform, Gmail, Google Sheets, Google Forms, Google Drive, Web
### Integration Pieces
-Connect your workflows to popular tools across the [Taskade integrations](https://www.taskade.com/integrations) directory. Premium integrations (for example Gmail, Slack, Google Sheets, HubSpot and Stripe) need a paid plan. The Free plan runs the built-in steps and the standard connectors (HTTP Request, RSS, Pushover and MCP Client).
+You can connect your workflows to the tools in the [Taskade integrations](https://www.taskade.com/integrations) directory. Premium integrations (for example Gmail, Slack, Google Sheets, HubSpot, and Stripe) need a paid plan. The Free plan runs the built-in steps and the standard connectors (HTTP Request, RSS, Pushover, and MCP Client).
| Category | Services |
| ---------------- | -------------------------------------------------------------------- |
@@ -125,20 +129,20 @@ Connect your workflows to popular tools across the [Taskade integrations](https:
| Developer | GitHub, HTTP |
| Forms & CMS | Typeform, Webflow, WordPress |
| Scheduling | Calendly, Cal.com |
-| AI & Media | Web Search, Web Scraping, YouTube Transcription, Document Extraction |
+| AI & Media | Search the web, Read a web page, Transcribe a YouTube video, Turn a file into text |
| Utilities | Schedule, Delay, Loop, Branch, Filter |
## Next steps
-* [Action & Trigger Reference](actions.md) — pick a trigger and add your first actions
-* [Automation Recipes](recipes.md) — start from a proven workflow
-* [Integration Overview](comprehensive-integration-guide.md) — connect the tools you already use
-* [Integration Options](integrations.md) — see the full list of connectors
-* [MCP Connectors](../genesis/mcp-connectors.md) — call any MCP server from an automation with the MCP Client step
-* [Living System Troubleshooting](../community-and-sharing/troubleshooting.md) — fix issues and get help
+* [Action & Trigger Reference](actions.md): select a trigger and add your first actions
+* [Automation Recipes](recipes.md): start from a workflow that works
+* [Integration Overview](comprehensive-integration-guide.md): connect the tools that you use now
+* [Integration Options](integrations.md): see the full list of connectors
+* [MCP Connectors](../genesis/mcp-connectors.md): call an MCP server from an automation with the MCP Client step
+* [Living System Troubleshooting](../community-and-sharing/troubleshooting.md): fix problems and get help
{% hint style="info" %}
-This page stays overview-level. Deep trigger/action lists, templates, and advanced patterns live in the reference pages above.
+This page gives an overview only. The reference pages above have the full lists of triggers and actions, the templates, and the advanced patterns.
{% endhint %}
---
diff --git a/genesis-living-system-builder/automation/actions.md b/genesis-living-system-builder/automation/actions.md
index 41ca6c4..6836733 100644
--- a/genesis-living-system-builder/automation/actions.md
+++ b/genesis-living-system-builder/automation/actions.md
@@ -5,14 +5,14 @@ description: >-
# Action & Trigger Reference
-Actions are the tasks that automations can perform. Each action connects to a specific service or system to create, update, or retrieve data. Actions can be chained together to create complex workflows.
+Actions are the steps that an automation does. Each action connects to a service or to your workspace. It creates, updates, or reads data. You can put actions one after the other to make a workflow.
## Overview
-Actions are the "then" part of automation workflows - the tasks that get performed when trigger conditions are met. Taskade provides a comprehensive set of actions that can create projects, manage tasks, interact with AI, integrate with external services, and perform complex data operations.
+A trigger starts an automation. The actions are the "then" part: they do the work when the trigger conditions are true. With actions, you can create projects, manage tasks, ask AI, send data to other apps, and change data.
{% hint style="success" %}
-Actions can be chained together to create complex multi-step workflows, with conditional logic and data transformation capabilities that enable truly intelligent automation.
+You can connect actions in a chain. Add a condition to choose a path, and transform data between steps.
{% endhint %}
## Complete Action Reference
@@ -21,59 +21,59 @@ Actions can be chained together to create complex multi-step workflows, with con
| Action | Description | Parameters |
| --- | --- | --- |
-| **Create Project** | Sets up a new project in your workspace/folder | `name`, `folderId`, `template`, `defaultView` |
-| **Add Task** | Adds a new task to an existing project | `projectId`, `title`, `content`, `assignee`, `dueDate` |
-| **Mark Task as Completed** | Automatically marks a task as done | `taskId`, `completedBy`, `completionNote` |
-| **Create Project From Template** | Uses a pre-defined or custom template | `templateId`, `name`, `folderId` |
+| **Create Project** | Creates a new project in your workspace or folder | `name`, `folderId`, `template`, `defaultView` |
+| **Add Task** | Adds a new task to a project | `projectId`, `title`, `content`, `assignee`, `dueDate` |
+| **Mark Task as Completed** | Marks a task as done | `taskId`, `completedBy`, `completionNote` |
+| **Create Project From Template** | Creates a project from a Taskade template or your own template | `templateId`, `name`, `folderId` |
| **Move Task** | Moves a task to another block in the same project | `projectId`, `taskId`, `targetBlockId` |
-| **Assign Task** | Assigns a task to someone | `taskId`, `userId`, `notifyAssignee` |
-| **Ask AI** | Uses AI to answer questions | `query`, `context`, `model` |
-| **Generate with AI** | Uses AI to generate content or responses | `prompt`, `style`, `length` |
-| **Respond with AI** | Automatically provides AI-generated responses | `input`, `tone`, `format` |
-| **Ask Agent** | Uses a custom AI agent to respond to a query | `agentId`, `query`, `parameters` |
-| **Run Agent Command** | Uses a custom AI agent command | `agentId`, `command`, `input` |
-| **Add Knowledge to Agent** | Transcribes knowledge and adds it as knowledge to an agent | `agentId`, `content`, `source` |
-| **Add Project to Agent Knowledge** | Automatically trains your AI agents with project content | `agentId`, `projectId`, `scope` |
-| **Ask Agent With Structured Output** | Structures the output of agent replies | `agentId`, `query`, `outputFormat` |
-| **Add a comment** | Appends a comment to an existing task node | `taskId`, `comment` |
-| **Update a task** | Updates a task node's content or note field | `taskId`, `content`, `note` |
-| **Get a task** | Reads a task node and returns its current data | `taskId` |
-| **Get project info** | Reads project metadata (name, dates, member list, etc.) | `projectId` |
-| **Attach a file** | Attaches a file to a task node (returns `attachmentId`, `fileId`, `projectId`, `nodeId`, `filename`, `mimetype`) | `projectId` (Project), `nodeId` (Task ID, UUID), `fileUrl` (File) |
+| **Assign Task** | Assigns a task to a person | `taskId`, `userId`, `notifyAssignee` |
+| **Ask AI** | Uses AI to answer a question | `query`, `context`, `model` |
+| **Write with AI** | Uses AI to write content or a reply | `prompt`, `style`, `length` |
+| **Reply with AI** | Writes a reply with AI | `input`, `tone`, `format` |
+| **Ask an agent** | Sends a question to one of your custom AI agents | `agentId`, `query`, `parameters` |
+| **Run an agent command** | Runs a saved command of a custom AI agent | `agentId`, `command`, `input` |
+| **Add knowledge to an agent** | Gives an agent a file, a website, a YouTube video, or text to learn from | `agentId`, `content`, `source` |
+| **Add a project to an agent** | Lets an agent read a project as knowledge | `agentId`, `projectId`, `scope` |
+| **Structured output** | Set **Output Type** to **Structured Output** on an **Ask an agent** or **Run an agent command** step. The step returns typed fields instead of free text | `outputType`, `output` |
+| **Add a comment** | Adds a comment to a task | `taskId`, `comment` |
+| **Update a task** | Changes the content or the note of a task | `taskId`, `content`, `note` |
+| **Get a task** | Reads a task and returns its current data | `taskId` |
+| **Get project info** | Reads the project data: name, dates, and members | `projectId` |
+| **Attach a file** | Attaches a file to a task. It returns `attachmentId`, `fileId`, `projectId`, `nodeId`, `filename`, and `mimetype` | `projectId` (Project), `nodeId` (Task ID, UUID), `fileUrl` (File) |
### Data Processing & Integration Actions
| Action | Description | Parameters |
| --- | --- | --- |
-| **Scrape Webpage** | Extract the contents of a web page | `url`, `errorBehavior` |
-| **Convert File to Text** | Convert a source file into text | `fileId`, `outputFormat`, `encoding` |
-| **Transcribe YouTube Video** | Turn a YouTube video into text | `youtubeUrl` |
-| **Turn a recording into a project** | Transcribes an audio recording into a new project | `audioRecording`, `projectTitle` |
-| **Upload File to Media** | Upload file through a URL to save in your workspace or folder's media tab | `url`, `fileName`, `folderId` |
-| **Send HTTP Request** | Makes an API request to an endpoint; authentication is optional per call | `method`, `url`, `headers`, `body`, optional auth |
-| **Search Web** | Searches the web for information to add to a project or for the AI to analyze | `query` |
+| **Read a web page** | Gets the content of a web page | `url`, `errorBehavior` |
+| **Turn a file into text** | Changes a source file into text | `fileId`, `outputFormat`, `encoding` |
+| **Transcribe a YouTube video** | Changes a YouTube video into text | `youtubeUrl` |
+| **Turn a recording into a project** | Writes an audio recording as text in a new project | `audioRecording`, `projectTitle` |
+| **Save a file to Media** | Downloads a file from a link into Media | `url`, `fileName`, `folderId` |
+| **Send HTTP Request** | Sends a request to an API endpoint. Authentication is optional for each call | `method`, `url`, `headers`, `body`, optional auth |
+| **Search the web** | Searches the web for information. You can add the result to a project or give it to AI | `query` |
{% hint style="info" %}
-**Scanned documents work too.** Convert File to Text now falls back to **vision OCR** when a file has no machine-readable text — scanned PDFs, image-only documents, and photos of pages are transcribed automatically. OCR runs count toward your workspace's AI usage.
+**Scanned documents work too.** If a file has no text that a computer can read, Turn a file into text uses **vision OCR**. OCR reads the words in scanned PDFs, image-only documents, and photos of pages. OCR runs count toward the AI usage of your workspace.
{% endhint %}
### Workflow Control Actions
| Action | Description | Parameters |
| --- | --- | --- |
-| **Loop** | Iterates data tasks | `dataSource`, `iterationLogic`, `maxIterations` |
-| **Filter Data** | Performs actions only when the data meets a certain condition | `condition`, `dataSource`, `filterCriteria` |
-| **Branch** | Creates different paths within your workflow, depending on certain conditions | `condition`, `truePath`, `falsePath` |
-| **Transform Array to String** | Converts the output of Find Task(s) and Find Row(s) actions into text | `arrayInput`, `template`, `separator` |
-| **Run Automation** | Runs a saved flow as a depth-guarded subflow (prevents infinite recursion) | `flowId`, `input` |
+| **Loop** | Runs the steps inside it once for each item | `dataSource`, `iterationLogic` |
+| **Filter Data** | Continues only when the data meets a condition | `condition`, `dataSource`, `filterCriteria` |
+| **Branch** | Makes different paths in your workflow. Each path has a condition | `condition`, `truePath`, `falsePath` |
+| **Transform Array to String** | Changes the output of Find Task(s) and Find Row(s) into text | `arrayInput`, `template`, `separator` |
+| **Run Automation** | Runs a saved automation as a subflow. A depth limit stops an endless chain | `flowId`, `input` |
### Advanced Actions
| Action | Description | Parameters |
| --- | --- | --- |
-| **Update Custom Fields** | Updates custom fields within projects | `taskId`, `fieldId`, `value`, `projectId` |
-| **Find Task(s)** | Searches for tasks within projects | `projectId`, `searchCriteria`, `maxResults` |
-| **Delay** | Adds a time delay to automation workflows | `duration`, `unit` |
+| **Update Custom Fields** | Changes the custom fields of a task | `taskId`, `fieldId`, `value`, `projectId` |
+| **Find Task(s)** | Finds tasks in projects | `projectId`, `searchCriteria`, `maxResults` |
+| **Delay** | Stops the automation for a time, then continues | `duration`, `unit` |
## Core Taskade Actions
@@ -81,30 +81,30 @@ Actions can be chained together to create complex multi-step workflows, with con
| Action | Description | Parameters |
| --- | --- | --- |
-| **Create Project** | Sets up a new project in your workspace/folder. Example: Create "Customer Feedback" project from template | `name`, `folderId`, `template`, `defaultView` |
-| **Create Project From Template** | Uses a pre-defined or custom template. Example: Create project from "Support Workflow" template | `templateId`, `name`, `folderId` |
-| **Add Task** | Adds a new task to an existing project. Example: Add task "Review feedback #123" | `projectId`, `title`, `content`, `assignee`, `dueDate` |
-| **Mark Task as Completed** | Automatically marks a task as done. Example: Mark task complete with resolution note | `taskId`, `completedBy`, `completionNote` |
-| **Move Task** | Moves a task to another block in the same project. Example: Move to the "Processed" block | `projectId`, `taskId`, `targetBlockId` |
-| **Assign Task** | Assigns a task to someone. Example: Assign to support specialist with notification | `taskId`, `userId`, `notifyAssignee` |
-| **Update Custom Fields** | Updates custom fields within projects. Example: Set priority to "High", status to "In Progress" | `taskId`, `fieldId`, `value`, `projectId` |
-| **Find Task(s)** | Searches for tasks within projects. Example: Find all tasks assigned to user | `projectId`, `searchCriteria`, `maxResults` |
+| **Create Project** | Creates a new project in your workspace or folder. Example: Create a "Customer Feedback" project from a template | `name`, `folderId`, `template`, `defaultView` |
+| **Create Project From Template** | Creates a project from a Taskade template or your own template. Example: Create a project from the "Support Workflow" template | `templateId`, `name`, `folderId` |
+| **Add Task** | Adds a new task to a project. Example: Add the task "Review feedback #123" | `projectId`, `title`, `content`, `assignee`, `dueDate` |
+| **Mark Task as Completed** | Marks a task as done. Example: Mark a task as done with a resolution note | `taskId`, `completedBy`, `completionNote` |
+| **Move Task** | Moves a task to another block in the same project. Example: Move the task to the "Processed" block | `projectId`, `taskId`, `targetBlockId` |
+| **Assign Task** | Assigns a task to a person. Example: Assign the task to a support specialist and send a notification | `taskId`, `userId`, `notifyAssignee` |
+| **Update Custom Fields** | Changes the custom fields of a task. Example: Set the priority to "High" and the status to "In Progress" | `taskId`, `fieldId`, `value`, `projectId` |
+| **Find Task(s)** | Finds tasks in projects. Example: Find all tasks of one person | `projectId`, `searchCriteria`, `maxResults` |
### Find Task(s) Automation Settings
{% hint style="info" %}
-The Find Task(s) action enables dynamic task retrieval within automation workflows. Search, filter, and process tasks based on specific criteria to create intelligent, data-driven automations.
+With the Find Task(s) action, an automation can get tasks while it runs. You can search and filter tasks by your criteria. Then later steps use the tasks that the action finds.
{% endhint %}
#### When to Use Find Tasks
-**Perfect for:**
+**Use it to:**
-* **Bulk processing** - Handle multiple tasks at once
-* **Conditional workflows** - Process tasks meeting specific criteria
-* **Reporting** - Generate summaries of task collections
-* **Data synchronization** - Keep external systems updated
-* **Quality control** - Identify tasks needing attention
+* **Process many tasks** - Do the same work on many tasks at one time
+* **Run conditional workflows** - Process only the tasks that meet your criteria
+* **Make reports** - Write summaries of a group of tasks
+* **Synchronize data** - Keep other apps up to date
+* **Control quality** - Find the tasks that need attention
#### Search Criteria Options
@@ -116,7 +116,7 @@ The Find Task(s) action enables dynamic task retrieval within automation workflo
| **Task Status** | Equals, Not Equals | Status | `"completed"` |
| **Assignee** | Equals, Not Equals, Empty, Not Empty | User | `"john.doe"` |
| **Due Date** | Before, After, Empty, Not Empty | Date | `today()` |
-| **Custom Fields** | Various conditions based on field type | Depends on field | `"high"` |
+| **Custom Fields** | The conditions depend on the field type | Depends on field | `"high"` |
#### Output Data Structure
@@ -141,13 +141,13 @@ The Find Task(s) action enables dynamic task retrieval within automation workflo
#### Find Tasks Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| --------------------------- | ------------------------------------------------------------- | --------------------------------- |
| **Overdue Task Alerts** | Schedule → Find Tasks (overdue) → Send Email | Fewer missed deadlines |
-| **Weekly Progress Reports** | Schedule → Find Tasks (completed this week) → Generate Report | Better project visibility |
-| **Workload Balancing** | Schedule → Find Tasks (by assignee) → Reassign Tasks | Optimized team productivity |
-| **Quality Assurance** | Schedule → Find Tasks (missing details) → Flag for Review | Improved task quality |
-| **Performance Analytics** | Schedule → Find Tasks (by criteria) → Update Dashboard | Data-driven insights |
+| **Weekly Progress Reports** | Schedule → Find Tasks (completed this week) → Generate Report | A clear view of project progress |
+| **Workload Balancing** | Schedule → Find Tasks (by assignee) → Reassign Tasks | Work divided fairly in the team |
+| **Quality Assurance** | Schedule → Find Tasks (missing details) → Flag for Review | Tasks with complete details |
+| **Performance Analytics** | Schedule → Find Tasks (by criteria) → Update Dashboard | Decisions based on data |
#### Advanced Filtering Examples
@@ -203,25 +203,25 @@ Find Tasks → Transform Array → Export to CSV
#### Best Practices
-* **Limit result sets** to prevent performance issues (max 1000 tasks)
-* **Use specific filters** to narrow down results efficiently
-* **Combine with sorting** for predictable processing order
-* **Handle empty results** gracefully in your workflows
-* **Test with small datasets** before scaling to large projects
+* **Limit the results** - Smaller results run faster
+* **Use specific filters** - Narrow the results to the tasks that you need
+* **Sort the results** - Then the automation processes the tasks in a known order
+* **Plan for no results** - Make sure that the workflow works when the action finds no task
+* **Test on a small project first** - Then use the automation on large projects
***
### Custom Fields Automation Settings
{% hint style="info" %}
-The Update Custom Fields action allows you to automatically update custom fields in projects, ensuring your task data stays current without manual updates.
+The Update Custom Fields action changes the custom fields of a task for you. Your task data stays current, and you do not update it by hand.
{% endhint %}
**Available Parameters:**
-* **PROJECT ID**: ID of the target project
-* **TASK ID**: ID of the task you want to update
-* **FIELDS**: Add fields that should be updated with their new values
+* **PROJECT ID**: The ID of the project
+* **TASK ID**: The ID of the task to update
+* **FIELDS**: The fields to update, each with its new value
**Example Usage:**
@@ -248,55 +248,55 @@ The Update Custom Fields action allows you to automatically update custom fields
| Action | Description | Parameters |
| --- | --- | --- |
-| **Ask AI** | Uses AI to answer questions. Example: "What's the priority of this ticket?" | `question`, `context`, `model` |
-| **Generate with AI** | Uses AI to generate content or responses. Example: Generate email response to customer | `prompt`, `contentType`, `length` |
-| **Respond with AI** | Automatically provides AI-generated responses. Example: Auto-respond to common inquiries | `input`, `responseType`, `tone` |
-| **Ask Agent** | Uses a custom AI agent to respond to a query. Example: Query specialized support agent | `agentId`, `question`, `context` |
-| **Run Agent Command** | Uses a custom AI agent command. Example: Run "analyze\_sentiment" command | `agentId`, `commandId`, `input` |
-| **Ask Agent With Structured Output** | Structures the output of agent replies. Example: Get JSON response with specific fields | `agentId`, `query`, `outputSchema` |
-| **Add Project to Agent Knowledge** | Automatically trains AI agents with project content. Example: Keep agents updated with latest project information | `agentId`, `projectId` |
+| **Ask AI** | Uses AI to answer a question. Example: "What is the priority of this ticket?" | `question`, `context`, `model` |
+| **Write with AI** | Uses AI to write content or a reply. Example: Write an email reply to a customer | `prompt`, `contentType`, `length` |
+| **Reply with AI** | Writes a reply with AI. Example: Reply to common questions | `input`, `responseType`, `tone` |
+| **Ask an agent** | Sends a question to one of your custom AI agents. Example: Ask a support agent | `agentId`, `question`, `context` |
+| **Run an agent command** | Runs a saved command of a custom AI agent. Example: Run the "analyze\_sentiment" command | `agentId`, `commandId`, `input` |
+| **Structured output** | Set **Output Type** to **Structured Output** on an **Ask an agent** or **Run an agent command** step. Example: Get typed fields back instead of free text | `outputType`, `output` |
+| **Add a project to an agent** | Lets an agent read a project as knowledge. Example: Keep an agent up to date with the newest project data | `agentId`, `projectId` |
{% hint style="info" %}
**Create and change agents from a flow:** The **Create an agent** action adds a new AI agent to the workspace with a name and instructions. The **Update an agent** action changes the name, the greeting, or the instructions of an agent.
{% endhint %}
-### Ask Agent With Structured Output Settings
+### Structured Output Settings
{% hint style="info" %}
-The Ask Agent With Structured Output action extracts specific information from AI agent responses and formats it as structured data that can be used as variables in subsequent automation steps.
+Structured output is a setting on the **Ask an agent** and **Run an agent command** steps, not a separate action. Set **Output Type** to **Structured Output**, then add the typed fields that you want back. Later steps can use each field as a variable.
-**BYOK support:** When a Bring Your Own Key (BYOK) model is configured on the target agent, the one-shot structured-output path honors that key, so inference runs against your own API account rather than Taskade's shared quota.
+**BYOK support:** You can give the agent a model on your own API key (BYOK). Then the structured output step uses your key, and your own API account pays for the call.
{% endhint %}
#### When to Use Structured Output
-**Perfect for:**
+**Use it to:**
-* Extracting specific data points from unstructured text
-* Converting AI responses into usable variables
-* Filling spreadsheet columns automatically
-* Creating conditional logic based on AI analysis
-* Standardizing data formats across automations
+* Get specific data points from text that has no fixed format
+* Change AI replies into variables that later steps can use
+* Fill spreadsheet columns for you
+* Choose a path in the workflow from the AI result
+* Use the same data format in all your automations
#### Output Schema Types
| Type | Description | Example Use Case |
| ----------------- | ------------------------------ | ------------------------------------------- |
-| **String** | Text data | Extract product names, titles, descriptions |
-| **Number** | Numeric values | Calculate costs, scores, percentages |
-| **Boolean** | True/false values | Check conditions, validate data |
-| **Date** | Date values | Extract deadlines, schedule dates |
-| **Single Select** | Choose from predefined options | Status: "pending", "approved", "rejected" |
+| **String** | Text data | Get product names, titles, descriptions |
+| **Number** | Numbers | Calculate costs, scores, percentages |
+| **Boolean** | True or false | Examine conditions, validate data |
+| **Date** | Dates | Get deadlines and schedule dates |
+| **Single Select** | One value from a list | Status: "pending", "approved", "rejected" |
#### Structured Output Use Cases
| Use Case | Automation Flow | Example |
| ----------------------- | --------------------------------------------- | ----------------------------------------- |
-| **Cost Analysis** | Task Added → Ask Agent → Update Fields | Calculate budget variance automatically |
-| **Lead Qualification** | Form Submitted → Ask Agent → Route Lead | Score and qualify incoming leads |
-| **Content Scoring** | Task Added → Ask Agent → Update Status | Rate content quality and assign reviewers |
-| **Performance Metrics** | Schedule → Ask Agent → Update Dashboard | Calculate completion rates and KPIs |
-| **Approval Routing** | Task Created → Ask Agent → Conditional Action | Route tasks based on AI evaluation |
+| **Cost Analysis** | Task Added → Ask an agent → Update Fields | Calculate the budget difference |
+| **Lead Qualification** | Form Submitted → Ask an agent → Route Lead | Score and qualify new leads |
+| **Content Scoring** | Task Added → Ask an agent → Update Status | Rate content quality and assign reviewers |
+| **Performance Metrics** | Schedule → Ask an agent → Update Dashboard | Calculate completion rates and KPIs |
+| **Approval Routing** | Task Created → Ask an agent → Conditional Action | Send tasks to a path from the AI result |
#### Example Schema Configuration
@@ -320,51 +320,51 @@ The Ask Agent With Structured Output action extracts specific information from A
}
```
-**Pro Tips:**
+**Tips:**
-* Use descriptive field names that match your workflow
-* Include clear descriptions for better AI understanding
-* Test schemas with sample data before deploying
-* Combine with conditional actions for powerful automation flows
+* Give each field a clear name that matches your workflow
+* Write a clear description for each field. The AI uses it to fill the field
+* Test the fields with sample data before you turn the automation on
+* Use the fields in a Branch or a Filter Data step
***
| Action | Description | Parameters |
| --- | --- | --- |
-| **Prompt AI** | Give the model detailed instructions. Example: Custom AI processing with specific instructions | `prompt`, `systemMessage`, `temperature` |
+| **Prompt AI** | Gives the model detailed instructions. Example: Process text with your own instructions | `prompt`, `systemMessage`, `temperature` |
-### Add Project to Agent Knowledge Automation Settings
+### Add a project to an agent Automation Settings
{% hint style="info" %}
-The Add Project to Agent Knowledge action automatically trains AI agents with project content, enabling context-aware responses and keeping agents updated with the latest information.
+The Add a project to an agent action lets an agent read a project as knowledge. The agent can then use the project details in its replies. The agent stays up to date with the newest project data.
{% endhint %}
-#### When to Use Add Project to Agent Knowledge
+#### When to Use Add a project to an agent
-**Perfect for:**
+**Use it to:**
-* **Training agents on new work** - Keep agents current with project information
-* **Enabling context-aware responses** - Allow agents to reference specific project details
-* **Supporting onboarding** - Feed previous projects to help new team members
-* **Powering research assistants** - Make agents smarter with research project content
-* **Preparing for launches** - Train agents on feature launch or campaign plans
+* **Teach agents about new work** - Give agents the newest project data
+* **Get replies with context** - Let agents use the details of a project
+* **Help new team members** - Give an agent past projects, so it can answer questions
+* **Make a research assistant** - Give an agent the content of research projects
+* **Prepare a launch** - Give an agent the plan of a feature launch or a campaign
-#### Add Project to Agent Knowledge Parameters
+#### Add a project to an agent Parameters
| Parameter | Description | Example |
| --------------------------- | ----------------------------------------------- | ---------------------------------- |
-| **Agent ID** | ID of the AI agent to train | `agent_123456` |
-| **Project ID** | ID of the project to add to knowledge | `proj_789012` |
+| **Agent ID** | The ID of the AI agent | `agent_123456` |
+| **Project ID** | The ID of the project to add as knowledge | `proj_789012` |
-#### Add Project to Agent Knowledge Use Cases
+#### Add a project to an agent Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| -------------------------- | ---------------------------------------------------------------- | ------------------------------------------ |
-| **New Project Onboarding** | Project Created → Add Project to Agent → Welcome Message | Instant agent training for new initiatives |
-| **Research Assistant** | Task Added → Add Project to Agent → Enable Queries | Context-aware research support |
-| **Customer Support** | Project Completed → Add Project to Agent → Update Knowledge Base | Enhanced support with project insights |
-| **Team Collaboration** | Task Added → Add Project to Agent → Notify Team | Improved team coordination |
-| **Knowledge Management** | Schedule → Add Project to Agent → Update Documentation | Automated knowledge base maintenance |
+| **New Project Onboarding** | Project Created → Add a project to an agent → Welcome Message | The agent knows each new project at once |
+| **Research Assistant** | Task Added → Add a project to an agent → Enable Queries | Research help with context |
+| **Customer Support** | Project Completed → Add a project to an agent → Update Knowledge Base | Support replies that use project data |
+| **Team Collaboration** | Task Added → Add a project to an agent → Notify Team | The team gets the same information |
+| **Knowledge Management** | Schedule → Add a project to an agent → Update Documentation | The knowledge base stays current |
#### Integration Examples
@@ -372,7 +372,7 @@ The Add Project to Agent Knowledge action automatically trains AI agents with pr
```
Trigger: Project Created
-Action: Add Project to Agent Knowledge
+Action: Add a project to an agent
Action: Send Channel Message (Agent trained on new project)
Result: Agents immediately knowledgeable about new projects
```
@@ -381,8 +381,8 @@ Result: Agents immediately knowledgeable about new projects
```
Trigger: Task Added (to the research project)
-Action: Add Project to Agent Knowledge
-Action: Ask Agent (Summarize new findings)
+Action: Add a project to an agent
+Action: Ask an agent (Summarize new findings)
Action: Update Dashboard (With insights)
Result: Real-time research insights and summaries
```
@@ -391,19 +391,19 @@ Result: Real-time research insights and summaries
```
Trigger: Task Added (To launch project)
-Action: Add Project to Agent Knowledge
-Action: Ask Agent (Generate FAQ)
+Action: Add a project to an agent
+Action: Ask an agent (Generate FAQ)
Action: Send Email (Team preparation)
Result: Comprehensive launch support and documentation
```
#### Best Practices
-* **Strategic Timing** - Add projects when they're most relevant to agent needs
-* **Content Filtering** - Add focused projects to avoid information overload
-* **Regular Updates** - Schedule periodic knowledge refreshes for evolving projects
-* **Agent Specialization** - Create dedicated agents for different project types
-* **Performance Monitoring** - Track agent response quality after knowledge updates
+* **Add projects at the right time** - Add a project when the agent needs it
+* **Add focused projects** - Too much data in one agent can make its replies less clear
+* **Refresh the knowledge** - Use a schedule to add a project again when it changes
+* **Use specialized agents** - Make one agent for each type of project
+* **Examine the replies** - Read the agent replies after you add new knowledge
***
@@ -411,40 +411,40 @@ Result: Comprehensive launch support and documentation
| Action | Description | Parameters |
| --- | --- | --- |
-| **Scrape Webpage** | Extract the contents of a web page. Example: Extract product info from competitor site | `url`, `errorBehavior` |
+| **Read a web page** | Gets the content of a web page. Example: Get product information from a public site | `url`, `errorBehavior` |
-### Scrape Webpage Automation Settings
+### Read a web page Automation Settings
{% hint style="info" %}
-The Scrape Webpage action enables automated data extraction from websites. Perfect for competitor monitoring, market research, lead generation, and price tracking without manual data collection.
+The Read a web page action gets data from a website for you. You can use it for market research, lead lists, and price tracking. You do not copy the data by hand.
{% endhint %}
-#### When to Use Scrape Webpage
+#### When to Use Read a web page
-**Perfect for:**
+**Use it to:**
-* **Competitor monitoring** - Track pricing, products, and promotions
-* **Market research** - Gather data from multiple sources automatically
-* **Lead generation** - Extract contact information from directories
-* **Price monitoring** - Track product prices across e-commerce sites
-* **Content aggregation** - Collect news, reviews, and industry updates
+* **Monitor other companies** - Track prices, products, and promotions
+* **Do market research** - Get data from many sources
+* **Find leads** - Get contact information from public directories
+* **Monitor prices** - Track product prices on store sites
+* **Collect content** - Collect news, reviews, and industry updates
-#### Scrape Webpage Parameters
+#### Read a web page Parameters
| Parameter | Description | Example |
| --------------- | ----------------------------------- | ------------------------------------- |
-| **URL** | Target webpage URL to scrape | `https://competitor.com/products` |
-| **Error Behavior** | What to do when the scrape fails: Continue or Fail | `Fail` (default) |
+| **URL** | The address of the web page | `https://competitor.com/products` |
+| **Error Behavior** | What the step does if it cannot read the page: Continue or Fail | `Fail` (default) |
-#### Scrape Webpage Use Cases
+#### Read a web page Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| ----------------------- | ---------------------------------------------------------------------- | ---------------------------------- |
-| **Competitor Analysis** | Schedule → Scrape Webpage → Analyze with AI → Update Dashboard | Real-time competitive intelligence |
-| **Price Monitoring** | Schedule → Scrape Webpage → Compare Prices → Send Alerts | Dynamic pricing optimization |
-| **Lead Generation** | Task Added → Scrape Webpage → Extract Contacts → Create Leads | Automated prospect discovery |
-| **Market Research** | Schedule → Scrape Webpage → Aggregate Data → Generate Reports | Comprehensive market insights |
-| **Content Monitoring** | Schedule → Scrape Webpage → Categorize Content → Update Knowledge Base | Current industry awareness |
+| **Competitor Analysis** | Schedule → Read a web page → Analyze with AI → Update Dashboard | Current data about the market |
+| **Price Monitoring** | Schedule → Read a web page → Compare Prices → Send Alerts | Fast reaction to price changes |
+| **Lead Generation** | Task Added → Read a web page → Extract Contacts → Create Leads | New leads without manual search |
+| **Market Research** | Schedule → Read a web page → Aggregate Data → Generate Reports | Regular market reports |
+| **Content Monitoring** | Schedule → Read a web page → Categorize Content → Update Knowledge Base | Current industry news |
#### Integration Examples
@@ -452,8 +452,8 @@ The Scrape Webpage action enables automated data extraction from websites. Perfe
```
Trigger: Schedule (Daily)
-Action: Scrape Webpage (Competitor Pricing Page)
-Action: Ask Agent With Structured Output (Extract and compare prices)
+Action: Read a web page (Competitor Pricing Page)
+Action: Ask an agent with structured output (Extract and compare prices)
Action: Send Slack Alert (If Significant Changes)
Result: Automated competitive pricing intelligence
```
@@ -462,8 +462,8 @@ Result: Automated competitive pricing intelligence
```
Trigger: Task Added (Industry Research)
-Action: Scrape Webpage (Industry Directory)
-Action: Ask Agent With Structured Output (Extract company details)
+Action: Read a web page (Industry Directory)
+Action: Ask an agent with structured output (Extract company details)
Action: Categorize with AI (Lead Qualification)
Action: Create Tasks (Follow-up Actions)
Result: Automated lead discovery and qualification
@@ -473,7 +473,7 @@ Result: Automated lead discovery and qualification
```
Trigger: Schedule (Hourly)
-Action: Scrape Webpage (Industry News Sites)
+Action: Read a web page (Industry News Sites)
Action: Categorize with AI (Document Classification)
Action: Add to Knowledge Base
Result: Current industry news awareness
@@ -481,7 +481,7 @@ Result: Current industry news awareness
#### Output Data Structure
-**Scraped Content Format:**
+**Page Content Format:**
```json
{
@@ -493,58 +493,58 @@ Result: Current industry news awareness
#### Best Practices
-* **Respect robots.txt** and website terms of service
-* **Use appropriate delays** between requests to avoid overloading servers
-* **Monitor scraping success** rates in the **Runs** tab
-* **Cache results** appropriately to reduce unnecessary requests
-* **Implement error handling** for network issues and site changes
+* **Obey robots.txt** and the terms of service of each website
+* **Do not send too many requests** to one server in a short time
+* **Monitor the results** in the **Runs** tab
+* **Keep results that you need again** - Then you send fewer requests
+* **Plan for errors** - A site can be down or change its layout
#### Legal and Ethical Considerations
-* **Terms of Service**: Review and comply with website terms
-* **Rate Limiting**: Avoid overwhelming target servers
-* **Data Usage**: Ensure compliance with data protection regulations
-* **Attribution**: Properly credit sources when using scraped content
-* **Fair Use**: Use scraped data responsibly and ethically
+* **Terms of Service**: Read the terms of each website and obey them
+* **Request Rate**: Do not send too many requests to a server
+* **Data Usage**: Obey the data protection laws that apply to you
+* **Attribution**: Name your sources when you use their content
+* **Fair Use**: Use the data in a responsible way
| Action | Description | Parameters |
| --- | --- | --- |
-| **Convert File to Text** | Convert a source file into text. Example: Convert PDF manual to searchable text | `fileUrl`, `fileType`, `extractImages` |
-| **Transcribe YouTube Video** | Turn a YouTube video into text. Example: Create transcript of training video | `youtubeUrl` |
-| **Turn a recording into a project** | Transcribes an audio recording into a new project. Example: Turn voice memo into project notes | `audioRecording`, `projectTitle` |
+| **Turn a file into text** | Changes a source file into text. Example: Change a PDF manual into text that you can search | `fileUrl`, `fileType`, `extractImages` |
+| **Transcribe a YouTube video** | Changes a YouTube video into text. Example: Make a transcript of a training video | `youtubeUrl` |
+| **Turn a recording into a project** | Writes an audio recording as text in a new project. Example: Turn a voice memo into project notes | `audioRecording`, `projectTitle` |
-### Transcribe YouTube Video Automation Settings
+### Transcribe a YouTube video Automation Settings
{% hint style="info" %}
-The Transcribe YouTube Video action converts video audio into searchable, readable text using advanced AI transcription. Perfect for content indexing, accessibility, educational resources, and content repurposing.
+The Transcribe a YouTube video action changes the audio of a video into text that you can read and search. You can use the text for an index, for accessibility, for study material, or for new content.
{% endhint %}
-#### When to Use Transcribe YouTube Video
+#### When to Use Transcribe a YouTube video
-**Perfect for:**
+**Use it to:**
-* **Content indexing** - Make video content searchable and referenceable
-* **Educational resources** - Create study guides and learning materials
-* **Content repurposing** - Transform video content into blogs, articles, and social posts
-* **Accessibility** - Provide text transcripts for hearing-impaired users
-* **Research and analysis** - Extract insights from video presentations
-* **Knowledge management** - Build searchable video content libraries
+* **Index content** - Make video content searchable
+* **Make study material** - Write study guides from a video
+* **Reuse content** - Change a video into a blog post, an article, or social posts
+* **Improve accessibility** - Give a text transcript to users who cannot hear the video
+* **Do research** - Get the main points of a video presentation
+* **Manage knowledge** - Make a library of video transcripts that you can search
-#### Transcribe YouTube Video Parameters
+#### Transcribe a YouTube video Parameters
| Parameter | Description | Example |
| -------------------------- | ----------------------------------------- | --------------------------------------------- |
-| **YouTube URL** | YouTube video URL to transcribe | `https://www.youtube.com/watch?v=VIDEO_ID` |
+| **YouTube URL** | The address of the YouTube video | `https://www.youtube.com/watch?v=VIDEO_ID` |
-#### Transcribe YouTube Video Use Cases
+#### Transcribe a YouTube video Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| ------------------------- | ---------------------------------------------- | --------------------------------------- |
| **Content Indexing** | Schedule → Transcribe → Index in Search | Searchable transcripts of your videos |
-| **Educational Resources** | Task Added → Transcribe → Generate Study Guide | Accelerated learning materials creation |
-| **Content Repurposing** | Video Upload → Transcribe → Generate Blog Post | Multi-platform content distribution |
-| **Accessibility** | Video Published → Transcribe → Add Captions | Improved content accessibility |
-| **Research Analysis** | Video Found → Transcribe → Extract Insights | Faster research and analysis |
+| **Educational Resources** | Task Added → Transcribe → Generate Study Guide | Faster study guides |
+| **Content Repurposing** | Video Upload → Transcribe → Generate Blog Post | The same content on more channels |
+| **Accessibility** | Video Published → Transcribe → Add Captions | Content that more people can use |
+| **Research Analysis** | Video Found → Transcribe → Extract Insights | Faster research |
#### Integration Examples
@@ -552,8 +552,8 @@ The Transcribe YouTube Video action converts video audio into searchable, readab
```
Trigger: Video Added to Course
-Action: Transcribe YouTube Video
-Action: Generate Study Guide with AI
+Action: Transcribe a YouTube video
+Action: Write with AI (Study guide)
Action: Create Quiz Questions
Action: Add to Learning Management System
Result: Automated course content creation
@@ -563,9 +563,9 @@ Result: Automated course content creation
```
Trigger: YouTube Video Published
-Action: Transcribe YouTube Video
-Action: Ask Agent With Structured Output (Extract key points)
-Action: Generate Blog Post with AI
+Action: Transcribe a YouTube video
+Action: Ask an agent with structured output (Extract key points)
+Action: Write with AI (Blog post)
Action: Publish to Website
Result: Multi-platform content distribution
```
@@ -574,8 +574,8 @@ Result: Multi-platform content distribution
```
Trigger: Weekly Schedule
-Action: Search Web (Industry videos)
-Action: Transcribe YouTube Video (Top results)
+Action: Search the web (Industry videos)
+Action: Transcribe a YouTube video (Top results)
Action: Categorize with AI (Topics)
Action: Generate Research Summary
Action: Add to Knowledge Base
@@ -584,20 +584,20 @@ Result: Automated industry research
#### Output Data Structure
-The action returns the video title, a metadata section, and a transcript section. Pass the transcript to an AI step to summarize it or pull out key points.
+The action returns the video title, a metadata section, and a transcript section. Give the transcript to an AI step to summarize it or to get the main points.
#### Best Practices
-* **Video Quality Matters** - Clear audio produces better transcripts
-* **Privacy Considerations** - Review content before public sharing
-* **Cost Management** - Long videos may incur higher processing costs
-* **Quality Validation** - Review important transcripts for accuracy
-* **Integration Planning** - Consider downstream processing of transcript data
+* **Use videos with clear audio** - Clear audio gives better transcripts
+* **Protect privacy** - Read the content before you share it in public
+* **Control cost** - Long videos can cost more to process
+* **Examine the quality** - Read important transcripts for errors
+* **Plan the next steps** - Decide how later steps use the transcript
-#### Rate Limiting & Performance
+#### Processing Time
-* **Processing Time** - Longer videos take more time to process
-* **Caching** - Store transcripts to avoid re-processing
+* **Processing Time** - A longer video takes more time to process
+* **Saved Transcripts** - Save transcripts, so you do not process the same video again
#### Content Repurposing Workflows
@@ -623,106 +623,106 @@ YouTube Video → Transcribe → Generate Quiz → Create Study Guide → Add to
| Action | Description | Parameters |
| --- | --- | --- |
-| **OpenAI Generate Image** | Create images from text prompts with your own OpenAI key. Example: Generate illustrations and visual content | `prompt`, `model`, `size` |
+| **OpenAI Generate Image** | Makes images from text prompts with your own OpenAI key. Example: Make illustrations for your content | `prompt`, `model`, `size` |
### HTTP Request Automation Settings
{% hint style="info" %}
-HTTP (HyperText Transfer Protocol) is the foundation of web communication. The HTTP Request action allows you to connect Taskade with any web service or API, enabling seamless data exchange and integration.
+With the HTTP Request action, Taskade can send data to almost any web service or API, and get data back. HTTP (HyperText Transfer Protocol) is the protocol that the web uses.
{% endhint %}
#### HTTP Request Parameters
| Parameter | Description | Example |
| ---------------------- | ------------------------------------------------ | ----------------------------------------------- |
-| **Method** | HTTP method (GET, POST, PUT, DELETE, etc.) | `POST` |
-| **URL** | The endpoint URL to send the request to | `https://api.example.com/v1/users` |
-| **Headers** | HTTP headers (Authorization, Content-Type, etc.) | `Authorization: Bearer token123` |
-| **Query Params** | URL query parameters | `?limit=10&sort=date` |
-| **Body Type** | Request body format (JSON, Form Data, Raw, XML) | `JSON` |
-| **Body Content** | The actual data to send | `{"name": "John", "email": "john@example.com"}` |
-| **Response Body Type** | How to process the response (JSON or None) | `JSON` |
+| **Method** | The HTTP method: GET, POST, PUT, DELETE, and more | `POST` |
+| **URL** | The endpoint address for the request | `https://api.example.com/v1/users` |
+| **Headers** | HTTP headers, for example Authorization and Content-Type | `Authorization: Bearer token123` |
+| **Query Params** | The query parameters in the URL | `?limit=10&sort=date` |
+| **Body Type** | The format of the request body: JSON, Form Data, Raw, or XML | `JSON` |
+| **Body Content** | The data to send | `{"name": "John", "email": "john@example.com"}` |
+| **Response Body Type** | How the step reads the response: JSON or None | `JSON` |
#### HTTP Request Use Cases
| Use Case | Automation Flow | Example |
| ----------------------- | --------------------------------------------- | ----------------------------------------- |
-| **Weather Data** | Schedule → HTTP Request → Add Task | Get daily weather for project planning |
-| **Email Validation** | Task Added → HTTP Request → Update Fields | Validate contact emails automatically |
-| **Content Translation** | Task Added → HTTP Request → Append Text | Translate content for international teams |
-| **Stock Prices** | Schedule → HTTP Request → Insert Row | Track investment portfolio data |
+| **Weather Data** | Schedule → HTTP Request → Add Task | Get the daily weather for project plans |
+| **Email Validation** | Task Added → HTTP Request → Update Fields | Validate contact emails |
+| **Content Translation** | Task Added → HTTP Request → Append Text | Translate content for teams in other countries |
+| **Stock Prices** | Schedule → HTTP Request → Insert Row | Track the data of your investments |
| **Public Holidays** | Project Created → HTTP Request → Create Event | Add holidays to project calendars |
-| **Business Lookup** | Form Submitted → HTTP Request → Update Fields | Enrich contact data with company info |
+| **Business Lookup** | Form Submitted → HTTP Request → Update Fields | Add company data to a contact |
### OpenAI Generate Image Settings
{% hint style="info" %}
-The OpenAI **Generate Image** step creates an image from a text prompt with your own OpenAI API key. OpenAI bills your account for the image, and the step uses a fixed step fee in Taskade credits. The step needs Pro and up. OpenAI retired its older DALL·E image models.
+The OpenAI **Generate Image** step makes an image from a text prompt with your own OpenAI API key. OpenAI bills your account for the image, and the step uses a fixed step fee in Taskade credits. The step needs Pro and up. OpenAI retired its older DALL·E image models.
{% endhint %}
#### Setup Requirements
**OpenAI API Key Required:**
-1. Visit [OpenAI Platform](https://platform.openai.com/api-keys) and create an API key
-2. In the automation, add a step and pick **OpenAI** → **Generate Image**
-3. Click **Connect** and paste the key. Taskade checks the key before it stores it
+1. Go to [OpenAI Platform](https://platform.openai.com/api-keys) and create an API key.
+2. In the automation, add a step and select **OpenAI** → **Generate Image**.
+3. Click **Connect** and paste the key. Taskade examines the key before it saves it.
#### Generate Image Parameters
| Parameter | Description | Options | Default |
| ---------- | ----------------------------------------- | ---------------------------------------- | --------------- |
-| **Prompt** | Text description of the image to generate | Any descriptive text | Required |
-| **Model** | Image model to use | Leave empty for the current default | Current default |
-| **Size** | Image dimensions | `1024x1024`, `1536x1024`, `1024x1536` | Optional |
+| **Prompt** | A text description of the image | Any descriptive text | Required |
+| **Model** | The image model | Leave empty for the current default | Current default |
+| **Size** | The size of the image in pixels | `1024x1024`, `1536x1024`, `1024x1536` | Optional |
-The step returns a hosted `url` for the image, plus `revisedPrompt` when the model returns one.
+The step returns a hosted `url` for the image. It also returns `revisedPrompt` when the model gives one.
#### Generate Image Use Cases
| Use Case | Automation Flow | Example |
| ----------------------------- | ------------------------------------------------ | ----------------------------------- |
-| **Visual Content Generation** | Task Added → Generate Image → Add to Project | Create illustrations for blog posts |
-| **Creative Project Support** | Task Added → Generate Image → Attach to Task | Add moodboards and concept images |
-| **Social Media Visuals** | Schedule → Generate Image → Post to Social | Create dynamic social media content |
-| **Learning Materials** | Task Added → Generate Image → Embed in Content | Make educational materials engaging |
-| **Product Prototyping** | Task Added → Generate Image → Review & Iterate | Produce concept images for products |
+| **Visual Content Generation** | Task Added → Generate Image → Add to Project | Make illustrations for blog posts |
+| **Creative Project Support** | Task Added → Generate Image → Attach to Task | Add mood boards and concept images |
+| **Social Media Visuals** | Schedule → Generate Image → Post to Social | Make images for social media |
+| **Learning Materials** | Task Added → Generate Image → Embed in Content | Add images to study materials |
+| **Product Prototyping** | Task Added → Generate Image → Review & Iterate | Make concept images for products |
-**Prompt Engineering Tips:**
+**Tips for prompts:**
-* Be specific about style, lighting, and composition
-* Specify image format and intended use
-* Use descriptive adjectives for better results
+* Describe the style, the light, and the composition
+* Say the image format and how you will use the image
+* Use clear adjectives
-For images without an API key, ask Taskade EVE or an agent to generate them. That path uses Taskade AI credits and saves the file to Media.
+To make images without an API key, ask Taskade EVE or an agent. That path uses Taskade AI credits and saves the file to Media.
### Taskade Genesis — Image Format Conversion
{% hint style="info" %}
-The **Convert an image** action converts an image from one format to another entirely within Taskade — no external service required. Useful for normalizing images before uploading them to downstream steps.
+The **Convert an image** action changes an image from one format to another in Taskade. It needs no outside service. Use it to give all images the same format before later steps use them.
{% endhint %}
| Action | Description | Parameters |
| --- | --- | --- |
-| **Convert an image** | Saves a copy of an image in another format: JPEG, PNG, WebP, or AVIF. Example: Convert an uploaded PNG to WebP before storing | `imageUrl`, `targetFormat` |
+| **Convert an image** | Saves a copy of an image in another format: JPEG, PNG, WebP, or AVIF. Example: Convert an uploaded PNG to WebP before you save it | `imageUrl`, `targetFormat` |
-The action strips EXIF metadata (GPS, camera info) by default. Output is returned as a file reference that can be passed to subsequent actions such as **Upload File to Media**.
+By default, the action removes EXIF metadata, for example the GPS location and the camera data. The output is a file reference. Later actions, such as **Save a file to Media**, can use it.
### Categorize with AI Automation Settings
{% hint style="info" %}
-The Categorize with AI action automatically analyzes, classifies, and organizes information using AI. Choose a preset or define a custom output schema for your specific needs.
+The Categorize with AI action uses AI to analyze, classify, and sort information. Select a preset, or make your own output schema.
{% endhint %}
#### When to Use Categorize with AI
-**Perfect for:**
+**Use it to:**
-* **Sentiment analysis** - Determine positive, neutral, or negative sentiment in feedback
-* **Lead qualification** - Score and prioritize potential customers
-* **Content moderation** - Filter inappropriate or sensitive content
-* **Support prioritization** - Route tickets based on urgency and type
-* **Data organization** - Automatically tag and sort incoming information
+* **Analyze sentiment** - Find out if feedback is positive, neutral, or negative
+* **Qualify leads** - Score and sort possible customers
+* **Moderate content** - Find content that is not correct for your community
+* **Prioritize support** - Send tickets to a path by urgency and type
+* **Organize data** - Tag and sort new information
#### Preset Categories Available
@@ -730,15 +730,26 @@ The Categorize with AI action automatically analyzes, classifies, and organizes
| --------------------------------- | ---------------------------------------- | ----------------------------------------------------------------- |
| **Sentiment Analysis** | Customer feedback, reviews, social media | Sentiment (Positive/Neutral/Negative), intensity, confidence, key phrases |
| **Lead Qualification** | Sales inquiries, prospect data | Qualification (Hot Lead/Warm Lead/Cold Lead), fit score, budget match |
-| **Content Moderation** | User-generated content, comments | Status (Approved/Flagged/Rejected), reason, risk level, action |
+| **Content Moderation** | User content, comments | Status (Approved/Flagged/Rejected), reason, risk level, action |
| **Support Ticket Prioritization** | Help tickets, customer issues | Priority (High/Medium/Low), category, response time, escalation level |
| **Document Classification** | Articles, documents | Type, quarter, year, relevance (Low/Medium/High) |
-More presets cover recruitment candidate scoring, customer feedback analysis, churn risk prediction, meeting agendas, marketing campaigns, project status, competitor analysis, expense categorization, email response prioritization, and social media engagement.
+More presets are available for these jobs:
+
+* Recruitment candidate scoring
+* Customer feedback analysis
+* Churn risk prediction
+* Meeting agendas
+* Marketing campaigns
+* Project status
+* Competitor analysis
+* Expense categorization
+* Email response prioritization
+* Social media engagement
#### Custom Category Configuration
-If no preset fits, define a custom output schema. Each field you add becomes a variable for later steps.
+If no preset is correct for your job, make a custom output schema. Each field that you add becomes a variable for later steps.
**Example: Customer Support Ticket Classification**
@@ -759,13 +770,13 @@ If no preset fits, define a custom output schema. Each field you add becomes a v
#### Categorize with AI Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| ---------------------- | ----------------------------------------------- | --------------------------------------- |
-| **Sentiment Analysis** | Form Submission → Categorize → Route Response | Faster, routed customer responses |
-| **Lead Qualification** | New Contact → Categorize → Assign Sales Rep | Leads routed to the right rep |
-| **Support Routing** | Ticket Created → Categorize → Assign Agent | Tickets categorized and assigned |
-| **Content Moderation** | User Post → Categorize → Approve/Reject | Maintains community standards |
-| **Feedback Analysis** | Survey Response → Categorize → Update Dashboard | Data-driven product improvements |
+| **Sentiment Analysis** | Form Submission → Categorize → Route Response | Faster replies to customers |
+| **Lead Qualification** | New Contact → Categorize → Assign Sales Rep | Each lead goes to the correct rep |
+| **Support Routing** | Ticket Created → Categorize → Assign Agent | Each ticket gets a category and an owner |
+| **Content Moderation** | User Post → Categorize → Approve/Reject | Your community rules stay in force |
+| **Feedback Analysis** | Survey Response → Categorize → Update Dashboard | Product decisions based on data |
#### Integration Examples
@@ -802,11 +813,11 @@ Branch: If "Approved" → Publish Immediately
#### Configuration Best Practices
-* **Start with presets** before creating custom categories
-* **Branch on the confidence field** where a preset returns one, to balance accuracy vs. coverage
-* **Test with sample data** to validate categorization accuracy
-* **Monitor and refine** categories based on real-world performance
-* **Combine with human review** for critical decisions
+* **Try the presets first** - Make custom categories only if no preset is correct
+* **Branch on the confidence field** - Some presets return a confidence value. Use it to decide which results need a person
+* **Test with sample data** - Make sure that the categories are correct
+* **Improve the categories** - Change them when real results show a problem
+* **Let a person decide important cases** - Add a review step for critical decisions
#### Output Data Structure
@@ -823,37 +834,37 @@ Branch: If "Approved" → Publish Immediately
***
-### Search Web Automation Settings
+### Search the web Automation Settings
{% hint style="info" %}
-The Search Web action enables automations to gather real-time information from the internet. Perfect for research, content creation, market analysis, and keeping your workflows updated with the latest information.
+With the Search the web action, an automation gets current information from the internet. You can use it for research, content, and market analysis. Your workflows then use the newest information.
{% endhint %}
-#### When to Use Search Web
+#### When to Use Search the web
-**Perfect for:**
+**Use it to:**
-* **Research automation** - Gather up-to-date information for reports and presentations
-* **Content creation** - Find relevant data and trends for blog posts and articles
-* **Market analysis** - Monitor competitor activity and industry trends
-* **Background checking** - Verify information and gather additional context
-* **Data enrichment** - Enhance existing data with web-sourced information
+* **Do research** - Get current information for reports and presentations
+* **Write content** - Find data and trends for blog posts and articles
+* **Analyze the market** - Monitor other companies and industry trends
+* **Examine information** - Make sure that information is correct, and get more context
+* **Add to your data** - Add information from the web to the data that you have
-#### Search Web Use Cases
+#### Search the web Use Cases
-| Use Case | Automation Flow | Business Impact |
+| Use Case | Automation Flow | Result |
| ------------------------- | ----------------------------------------- | -------------------------- |
-| **Resume Verification** | Task Added → Search Web → Compare Results | Improved hiring quality |
-| **Content Research** | Schedule → Search Web → Generate Article | Faster content creation |
-| **Market Intelligence** | Schedule → Search Web → Update Dashboard | Better strategic decisions |
-| **Competitor Analysis** | Task Added → Search Web → Analyze Trends | Competitive advantage |
-| **Automated Newsletters** | Schedule → Search Web → Curate Content | Engaging communications |
+| **Resume Verification** | Task Added → Search the web → Compare Results | Better hiring decisions |
+| **Content Research** | Schedule → Search the web → Generate Article | Faster content |
+| **Market Intelligence** | Schedule → Search the web → Update Dashboard | Better plans |
+| **Competitor Analysis** | Task Added → Search the web → Analyze Trends | Current market knowledge |
+| **Automated Newsletters** | Schedule → Search the web → Curate Content | Regular newsletters |
#### Configuration Options
| Setting | Description | Example |
| --------- | ---------------------------------- | ----------------------- |
-| **Query** | Search terms to use for web search | "latest AI trends 2026" |
+| **Query** | The words to search for | "latest AI trends 2026" |
#### Integration Examples
@@ -861,8 +872,8 @@ The Search Web action enables automations to gather real-time information from t
```
Trigger: Task Added (Research Topic)
-Action: Search Web (Query from task name)
-Action: Ask Agent With Structured Output (Summarize findings)
+Action: Search the web (Query from task name)
+Action: Ask an agent with structured output (Summarize findings)
Action: Create Project (Research Report)
Result: Automated research compilation
```
@@ -871,9 +882,9 @@ Result: Automated research compilation
```
Trigger: Schedule (Daily)
-Action: Search Web (Industry news)
+Action: Search the web (Industry news)
Action: Categorize with AI (Topic classification)
-Action: Generate with AI (News summary)
+Action: Write with AI (News summary)
Action: Publish to Blog
Result: Automated content publishing
```
@@ -882,7 +893,7 @@ Result: Automated content publishing
```
Trigger: Schedule (Weekly)
-Action: Search Web (Competitor name + news)
+Action: Search the web (Competitor name + news)
Action: Analyze with AI (Impact assessment)
Action: Update Custom Fields (Competitor activity)
Action: Send Slack Alert (If significant changes)
@@ -891,7 +902,7 @@ Result: Proactive competitive intelligence
#### Output Data Structure
-The action returns an AI-generated summary of the top results:
+The action returns an AI summary of the top results:
```json
{
@@ -901,18 +912,16 @@ The action returns an AI-generated summary of the top results:
#### Best Practices
-* **Craft specific queries** for better, more relevant results
-* **Name the time frame in the query** to focus on recent information
-* **Combine with AI analysis** to process and summarize findings
-* **Set reasonable limits** to avoid overwhelming downstream actions
-* **Monitor search costs** for high-volume automations
+* **Write specific queries** - Specific words give better results
+* **Name the time frame in the query** - Then you get recent information
+* **Add an AI step** - Use AI to process and summarize the results
+* **Keep the results small** - Too much data can slow later steps
+* **Monitor the cost** - Examine the cost of automations that search many times
-#### Rate Limiting & Performance
+#### Performance
-* **API quotas** may apply depending on search provider
-* **Response times** vary based on query complexity
-* **Caching** can be implemented for repeated searches
-* **Error handling** important for network-dependent actions
+* **Response times** - A complex query can take more time
+* **Error handling** - The action needs the network. Plan for a failed search
***
@@ -926,52 +935,52 @@ Most third-party connectors below are premium integrations, so they need a paid
| Action | Description | Example |
| ---------------------------- | ------------------------------------------------------------------ | ------------------------------- |
-| **Send Channel Message** | Posts a message to a selected channel | Notify team of new ticket |
-| **Send Direct Message** | Sends a personal message to a user | Alert assignee of urgent task |
-| **Create Channel** | Creates a new Slack channel | Create project-specific channel |
-| **Add User to Channel** | Adds a user to a channel | Add stakeholder to discussion |
-| **Set User's Current Slack Status** | Sets status text and emoji | Set "In meeting" status |
-| **Schedule Channel Message** | Schedules a message to post to a channel later, up to 120 days ahead | Send reminder tomorrow |
+| **Send Channel Message** | Sends a message to a channel | Tell the team about a new ticket |
+| **Send Direct Message** | Sends a personal message to a user | Tell the assignee about an urgent task |
+| **Create Channel** | Creates a new Slack channel | Create a channel for a project |
+| **Add User to Channel** | Adds a user to a channel | Add a stakeholder to a discussion |
+| **Set User's Current Slack Status** | Sets the status text and emoji | Set the "In meeting" status |
+| **Schedule Channel Message** | Sends a message to a channel later, up to 120 days ahead | Send a reminder tomorrow |
| **Find User by Handle** / **Find User by Email** | Finds a user by Slack handle or email address | Match a form email to a user |
-| **List Thread Replies** | Reads the parent message of a thread and its replies | Summarize a thread with AI |
-| **Fetch Channel Messages** | Reads recent messages from a channel | Build a daily digest |
-| **Add Reaction to Message** / **Pin Message to Channel** | Reacts to or pins a message | Pin the release note |
-| **Request Approval in a Channel** | Posts an approval request and waits for a reply | Approve a post before it goes out |
+| **List Thread Replies** | Reads the first message of a thread and its replies | Summarize a thread with AI |
+| **Fetch Channel Messages** | Reads the recent messages of a channel | Make a daily digest |
+| **Add Reaction to Message** / **Pin Message to Channel** | Adds a reaction to a message, or pins it | Pin the release note |
+| **Request Approval in a Channel** | Sends an approval request and waits for a reply | Approve a post before it goes out |
### Email
| Action | Description | Example |
| ----------------------------------- | --------------------------------------------------- | --------------------------------- |
-| **Gmail: Send Email** | Sends an email through Gmail to a recipient | Send confirmation to customer |
-| **Gmail: Reply to Email** / **Forward Email** | Replies to or forwards an email | Reply to a support request |
+| **Gmail: Send Email** | Sends an email through Gmail | Send a confirmation to a customer |
+| **Gmail: Reply to Email** / **Forward Email** | Replies to an email, or forwards it | Reply to a support request |
| **Gmail: Find Email(s)** | Finds emails in your Gmail account | Find the last order email |
-| **Gmail: Create Draft Email** | Creates a new draft email message | Draft a reply for review |
-| **Gmail: Get Thread** | Gets every message of an email thread, oldest first | Summarize a thread |
-| **Gmail: Create Draft Reply** | Saves a reply in an existing thread as a draft | Let a person review an AI reply |
+| **Gmail: Create Draft Email** | Creates a new draft email | Write a draft reply for review |
+| **Gmail: Get Thread** | Gets all messages of an email thread, oldest first | Summarize a thread |
+| **Gmail: Create Draft Reply** | Saves a reply in a thread as a draft | Let a person examine an AI reply |
| **Gmail: Request Approval by Email** | Sends an approval request by email | Approve a refund |
-| **Mailchimp: Add Contact** | Adds a contact to a Mailchimp audience | Subscribe to newsletter |
-| **Mailchimp: Update Contact Status** | Changes the subscription status of a contact | Unsubscribe user |
+| **Mailchimp: Add Contact** | Adds a contact to a Mailchimp audience | Add a subscriber to a newsletter |
+| **Mailchimp: Update Contact Status** | Changes the subscription status of a contact | Unsubscribe a user |
### Discord
| Action | Description | Example |
| -------------------------- | ------------------------------------- | ------------------------------- |
-| **Send Message** | Sends a message to a Discord channel | Post update to community |
-| **Create Guild Role** | Creates a new role | Create "Beta Tester" role |
-| **Add Guild Member Role** | Assigns a role to a member | Grant access to private channel |
+| **Send Message** | Sends a message to a Discord channel | Send an update to the community |
+| **Create Guild Role** | Creates a new role | Create a "Beta Tester" role |
+| **Add Guild Member Role** | Gives a role to a member | Give access to a private channel |
| **Create Channel** / **Rename Channel** / **Delete Channel** | Manages channels | Create a channel for each event |
### Notify by Email (Native)
{% hint style="success" %}
-Taskade includes a built-in **Send an email** action that sends notifications through Taskade's own mail infrastructure. No OAuth connection or third-party email provider is required.
+Taskade has a built-in **Send an email** action. It sends notifications through the Taskade mail service. You do not need an OAuth connection or another email provider.
{% endhint %}
| Action | Description | Parameters |
| --- | --- | --- |
-| **Send an email** | Sends an email from Taskade's verified address to you or a verified member of this workspace. Example: Alert a teammate when a task is created | `to`, `subject`, `message` |
+| **Send an email** | Sends an email from the verified Taskade address to you or a verified member of this workspace. Example: Tell a teammate about a new task | `to`, `subject`, `message` |
-To email anyone else, use the Gmail or Outlook send email action. If the step does not send to a recipient, the run history shows the step as **Skipped** with the reason, and the automation keeps running.
+To email another person, use the Gmail or Outlook send email action. If the step does not send to a recipient, the run history shows the step as **Skipped** with the reason. The automation continues.
## Data & CRM Actions
@@ -979,41 +988,41 @@ To email anyone else, use the Gmail or Outlook send email action. If the step do
| Action | Description | Example |
| ----------------------------- | ------------------------------------ | ------------------------ |
-| **Create New Contact** | Creates a new contact in HubSpot CRM | Add lead from form |
-| **Create or Update Contact** | Creates a contact, or updates it if it exists | Upsert a lead |
-| **Update Contact** | Updates contact properties | Update lead score |
-| **Create New Deal** | Creates a new deal in HubSpot CRM | Create sales opportunity |
-| **Update Deal** | Updates deal properties | Move to next stage |
-| **Find Contact** / **Find Deal** / **Search Contacts** / **Search Deals** | Looks up records | Check for an existing lead |
+| **Create New Contact** | Creates a new contact in HubSpot CRM | Add a lead from a form |
+| **Create or Update Contact** | Creates a contact, or updates it if it exists | Add or update a lead |
+| **Update Contact** | Changes the properties of a contact | Update a lead score |
+| **Create New Deal** | Creates a new deal in HubSpot CRM | Create a sales opportunity |
+| **Update Deal** | Changes the properties of a deal | Move a deal to the next stage |
+| **Find Contact** / **Find Deal** / **Search Contacts** / **Search Deals** | Finds records | Find an existing lead |
### Salesforce
| Action | Description | Example |
| ------------------------- | --------------------------------------------------------------- | ----------------------- |
-| **Create Record** | Creates a new record (Lead, Contact, Account, and more) | Add qualified lead |
-| **Update Record** | Updates an existing Salesforce record | Update lead status |
-| **Find Records by Query** | Searches for Salesforce records with a SOQL query | Find a customer record |
+| **Create Record** | Creates a new record (Lead, Contact, Account, and more) | Add a qualified lead |
+| **Update Record** | Changes a Salesforce record | Update a lead status |
+| **Find Records by Query** | Finds Salesforce records with a SOQL query | Find a customer record |
### Google Sheets
| Action | Description | Example |
| --------------- | ---------------------------------------------------------------- | ------------------------ |
-| **Insert Row** | Adds a new row to a sheet | Log form submission |
-| **Add or Update Multiple Rows** | Writes several rows in one step | Sync a list of orders |
-| **Update Row** | Updates a row in a sheet | Update status column |
-| **Find Row(s)** | Retrieves rows that match defined criteria (works with loops) | Find customer record |
+| **Insert Row** | Adds a new row to a sheet | Save a form submission |
+| **Add or Update Multiple Rows** | Writes many rows in one step | Sync a list of orders |
+| **Update Row** | Changes a row in a sheet | Update a status column |
+| **Find Row(s)** | Gets the rows that match your criteria. It works with loops | Find a customer record |
| **Get Rows** / **Get Row by Row Number** | Reads rows from a sheet | Load a price list |
| **Read Range** / **Clear Range** | Reads or clears the cells of a range in A1 notation | Read `A1:D20` |
-| **Delete Row** | Deletes a row in a sheet | Remove processed entry |
+| **Delete Row** | Deletes a row in a sheet | Remove a processed entry |
| **Create Spreadsheet** / **Create Worksheet** | Creates a new spreadsheet or tab | Start a monthly report |
### Airtable
| Action | Description | Example |
| ----------------- | ----------------------------------------------- | --------------------- |
-| **Create Record** | Creates a new record in an Airtable base | Add customer to CRM |
-| **List Records** | Lists records from a base with optional filters | Find open orders |
-| **Update Record** | Updates an existing Airtable record | Update project status |
+| **Create Record** | Creates a new record in an Airtable base | Add a customer to a CRM |
+| **List Records** | Lists the records of a base, with optional filters | Find open orders |
+| **Update Record** | Changes an Airtable record | Update a project status |
## Development Actions
@@ -1021,21 +1030,21 @@ To email anyone else, use the Gmail or Outlook send email action. If the step do
| Action | Description | Example |
| ------------------------- | ----------------------------------- | ------------------------ |
-| **Create Issue** | Opens a new issue in a repository | Report bug from feedback |
-| **Update Issue** | Updates an issue | Close resolved issue |
+| **Create Issue** | Opens a new issue in a repository | Report a bug from feedback |
+| **Update Issue** | Changes an issue | Close a resolved issue |
| **Create Comment** | Adds a comment to an issue or pull request | Add investigation notes |
-| **Get Issue Information** | Retrieves details about an issue | Read issue status |
-| **Lock Issue** / **Unlock Issue** | Locks or unlocks comments on an issue | Lock a resolved thread |
+| **Get Issue Information** | Gets the details of an issue | Read the issue status |
+| **Lock Issue** / **Unlock Issue** | Locks or unlocks the comments of an issue | Lock a resolved thread |
### Jira
| Action | Description | Example |
| ----------------------- | ------------------------------------- | --------------------------- |
-| **Create Issue** | Creates a new Jira issue | Create bug ticket |
-| **Update Issue** | Updates issue fields | Update priority or assignee |
+| **Create Issue** | Creates a new Jira issue | Create a bug ticket |
+| **Update Issue** | Changes the fields of an issue | Update the priority or assignee |
| **Move Issue to Status** | Changes the issue status | Move to "In Progress" |
| **Add Comment** | Adds a comment | Add troubleshooting notes |
-| **Get Issue** / **Search Issues** | Reads an issue or searches with JQL | Find open bugs |
+| **Get Issue** / **Search Issues** | Reads an issue, or searches with JQL | Find open bugs |
## Web & Content Actions
@@ -1043,18 +1052,18 @@ To email anyone else, use the Gmail or Outlook send email action. If the step do
| Action | Description | Example |
| --------------- | --------------------------------------------------------------------------- | -------------------- |
-| **Create Post** | Publishes a new post, or saves it as a draft. Add a Featured Image URL to set the featured image | Publish blog article |
+| **Create Post** | Publishes a new post, or saves it as a draft. Add a Featured Image URL to set the featured image | Publish a blog article |
### Webflow
| Action | Description | Example |
| -------------------- | -------------------------------------------- | -------------------- |
-| **List Sites** | Lists the Webflow sites available to the connection | Pick a site |
-| **List Collections** | Lists CMS collections for a selected site | Pick a collection |
-| **Create CMS Item** | Creates a staged collection item | Add new blog post |
-| **Get CMS Item** | Retrieves an item and its fields | Read product details |
-| **Update CMS Item** | Changes fields on an existing item | Edit product details |
-| **Publish CMS Item** | Publishes a staged item to the live site | Deploy updates live |
+| **List Sites** | Lists the Webflow sites of the connection | Select a site |
+| **List Collections** | Lists the CMS collections of a site | Select a collection |
+| **Create CMS Item** | Creates a staged collection item | Add a new blog post |
+| **Get CMS Item** | Gets an item and its fields | Read product details |
+| **Update CMS Item** | Changes the fields of an item | Edit product details |
+| **Publish CMS Item** | Publishes a staged item to the live site | Make updates live |
## Utility Actions
@@ -1062,51 +1071,51 @@ To email anyone else, use the Gmail or Outlook send email action. If the step do
| Action | Description | Parameters |
| --- | --- | --- |
-| **Send HTTP Request** | Calls a web service with GET, POST, PUT, PATCH, DELETE, or HEAD. Example: Call external API | `method`, `url`, `headers`, `queryParams`, `body` |
-| **Webhook** (trigger) | Starts the automation when an outside service posts to its private address. Needs Pro and up. Example: Receive a payment event | — |
+| **Send HTTP Request** | Calls a web service with GET, POST, PUT, PATCH, DELETE, or HEAD. Example: Call an outside API | `method`, `url`, `headers`, `queryParams`, `body` |
+| **Webhook** (trigger) | Starts the automation when an outside service sends data to its private address. Needs Pro and up. Example: Receive a payment event | — |
### Data Processing
-These actions work on text and lists without an AI call.
+These actions change text and lists without AI.
| Action | Description | Example |
| ------------------------------- | ------------------------------------------------------------ | ------------------------------------ |
-| **Split Text into a List** | Splits a text string into a list on a delimiter | Split CSV line into fields |
-| **Trim Extra Spaces** | Removes leading and trailing whitespace from a text string | Clean user-submitted input |
-| **Change Capitalization** | Converts text case | Normalize status labels to lowercase |
-| **Merge Two Sets of Data** | Combines two sets of data and picks which one wins on a shared field | Combine multiple sources |
-| **Get Multiple Values (JSONPath)** | Pulls values out of the data an earlier step returned | Read a nested field |
+| **Split Text into a List** | Divides text into a list at a delimiter | Divide a CSV line into fields |
+| **Trim Extra Spaces** | Removes the spaces at the start and at the end of text | Clean user input |
+| **Change Capitalization** | Changes the case of text | Make status labels lowercase |
+| **Merge Two Sets of Data** | Combines two sets of data. You select which set wins on a shared field | Combine two sources |
+| **Get Multiple Values (JSONPath)** | Gets values from the data that an earlier step returned | Read a nested field |
### Utilities
-Pure, no-credit transform actions for reshaping data inline in a flow — no AI call required.
+These actions change data in the flow. They do not call AI and use no credits.
| Action | Description | Parameters |
| --- | --- | --- |
-| **Generate QR Code** | Renders text or a URL into a QR image saved to your workspace Media library. Example: Encode a check-in URL as a QR code | `text` (required), `size` (150/300/600 px), `errorCorrection` (L/M/Q/H) |
-| **Convert CSV to JSON** | Parses a CSV string into structured rows. Example: Turn an upload into JSON rows | `csv` (required), `hasHeader` (header/noHeader), `delimiter` (optional, auto-detect) |
-| **Convert JSON to CSV** | Serializes a JSON array into CSV. Example: Export records to a spreadsheet | `json` (required), `includeHeader` (yes/no), `delimiter` (optional, default `,`) |
-| **Average** | Mean of a list of numbers. Example: Average response time | `values` (required) |
-| **Sum** | Total of a list of numbers. Example: Sum order amounts | `values` (required) |
-| **Count Unique Values** | Counts the distinct values in a list. Example: Count unique customers | `values` (required) |
-| **Min / Max** | Highest or lowest value in a list. Example: Find the largest invoice | `values` (required), `mode` (max/min) |
-| **Concatenate Text** | Joins multiple parts into one string. Example: Build a full name from fields | `parts` (required), `separator` (optional) |
-| **Find Text** | Locates a search value in a string (first match or all). Example: Find a keyword in a message | `text` (required), `searchValue` (required), `mode` (first/all), `caseSensitivity` (optional) |
-| **Slugify Text** | Turns text into a URL-safe slug. Example: Slugify an article title | `text` (required), `casing` (optional), `separator` (default `-`) |
-| **Strip HTML** | Removes markup and returns plain text. Example: Clean HTML email body | `html` (required) |
-| **HTML to Markdown** | Converts an HTML fragment to Markdown. Example: Store content as Markdown | `html` (required) |
-| **Markdown to HTML** | Renders Markdown to sanitized HTML. Example: Publish formatted content | `markdown` (required) |
+| **Generate QR Code** | Makes a QR image from text or a URL and saves it to your workspace Media library. Example: Make a QR code for a check-in URL | `text` (required), `size` (150/300/600 px), `errorCorrection` (L/M/Q/H) |
+| **Convert CSV to JSON** | Changes a CSV string into rows of data. Example: Turn an upload into JSON rows | `csv` (required), `hasHeader` (header/noHeader), `delimiter` (optional, auto-detect) |
+| **Convert JSON to CSV** | Changes a JSON array into CSV. Example: Export records to a spreadsheet | `json` (required), `includeHeader` (yes/no), `delimiter` (optional, default `,`) |
+| **Average** | The mean of a list of numbers. Example: Average response time | `values` (required) |
+| **Sum** | The total of a list of numbers. Example: Sum order amounts | `values` (required) |
+| **Count Unique Values** | Counts the different values in a list. Example: Count unique customers | `values` (required) |
+| **Min / Max** | The highest or lowest value in a list. Example: Find the largest invoice | `values` (required), `mode` (max/min) |
+| **Concatenate Text** | Joins many parts into one string. Example: Make a full name from fields | `parts` (required), `separator` (optional) |
+| **Find Text** | Finds a value in a string (the first match or all matches). Example: Find a keyword in a message | `text` (required), `searchValue` (required), `mode` (first/all), `caseSensitivity` (optional) |
+| **Slugify Text** | Changes text into a slug that is safe in a URL. Example: Make a slug from an article title | `text` (required), `casing` (optional), `separator` (default `-`) |
+| **Strip HTML** | Removes markup and returns plain text. Example: Clean an HTML email body | `html` (required) |
+| **HTML to Markdown** | Changes HTML into Markdown. Example: Save content as Markdown | `html` (required) |
+| **Markdown to HTML** | Changes Markdown into safe HTML. Example: Publish formatted content | `markdown` (required) |
### File Operations
| Action | Description | Example |
| ---------------------------- | ------------------------------------------------------------ | ---------------------- |
-| **Upload File to Media** | Uploads a file from a URL to the Media tab | Store form attachment |
+| **Save a file to Media** | Downloads a file from a link into Media | Save a form attachment |
| **Save text to Media** | Saves text as a file in the Media tab | Save an AI report |
-| **Find Files in Media** | Lists files by folder, file type, and name | Find this week's uploads |
+| **Find Files in Media** | Lists files by folder, file type, and name | Find the uploads of this week |
| **Convert an image** | Saves a copy of an image as JPEG, PNG, WebP, or AVIF | PNG to WebP |
-| **Make a smaller video** | Saves a smaller copy of a video | Reduce file size |
-| **Convert File to Text** | Converts a source file into text | PDF to text |
+| **Make a smaller video** | Saves a smaller copy of a video | Make a file smaller |
+| **Turn a file into text** | Changes a source file into text | PDF to text |
## Conditional Actions
@@ -1114,21 +1123,21 @@ Pure, no-credit transform actions for reshaping data inline in a flow — no AI
| Action | Description | Example |
| ---------------------------- | ----------------------------------------------------------------- | ----------------------------------- |
-| **Branch** | Creates different paths depending on conditions | If priority is high, notify manager |
+| **Branch** | Makes different paths. Each path has a condition | If the priority is high, tell the manager |
| **Filter Data** | Continues only when the data meets a condition | Skip low-value leads |
| **Loop** | Runs the steps inside it once for each item in a list | Process each item |
-| **Delay** | Pauses the automation for a set time or until a date | Wait 5 minutes before the next step |
-| **Run Automation** | Starts another automation from this one | Reuse a shared sub-flow |
+| **Delay** | Stops the automation for a set time or until a date | Wait 5 minutes before the next step |
+| **Run Automation** | Starts another automation from this one | Use a shared subflow again |
## Error Handling
| Action or setting | Description | Example |
| ------------------------------- | --------------------------------------------------------------------------- | ----------------------- |
-| **Stop and mark as failed** | Stops a run on purpose and marks it as failed | Stop on missing data |
-| **Get run info** | Reads the ID and the start time of the current run | Log the run ID |
-| **Error Behavior** (step setting) | On steps that offer it, choose **Continue** or **Fail** when the step errors | Keep going after a failed scrape |
+| **Stop and mark as failed** | Stops a run on purpose and marks it as failed | Stop when data is missing |
+| **Get run info** | Reads the ID and the start time of the current run | Save the run ID |
+| **Error Behavior** (step setting) | On steps that have it, select **Continue** or **Fail** for when the step has an error | Continue after a page cannot be read |
-Every run is listed in the automation **Runs** tab, with the input and output of each step.
+The **Runs** tab of the automation lists every run, with the input and the output of each step.
## Usage Examples
@@ -1165,11 +1174,11 @@ Branch: If the response is missing a required field
## **Advanced File Handling**
-Recent updates have significantly enhanced automation file processing capabilities.
+Automations can send, receive, and process files.
### **Multipart File Uploads**
-HTTP actions now support multipart file uploads, enabling seamless integration with modern APIs:
+HTTP actions can send files as a multipart upload. Many modern APIs accept files in this format:
```json
{
@@ -1190,23 +1199,23 @@ HTTP actions now support multipart file uploads, enabling seamless integration w
}
```
-### **Enhanced Webhook File Processing**
+### **Webhook File Processing**
-Webhooks can now receive and process multipart/form-data, perfect for:
+Webhooks can receive and process multipart/form-data. Use this for:
-* **Document Processing Workflows**: Automatically process uploaded files
-* **Image Analysis Pipelines**: Analyze images submitted through forms
-* **File Validation Systems**: Check file types and sizes before processing
+* **Document workflows**: Process uploaded files
+* **Image analysis**: Analyze images that people send through forms
+* **File validation**: Examine the file type and size before later steps
-### **Smart Form Field Mapping**
+### **Form Field Mapping**
-When forms trigger automations that create tasks, the system intelligently maps form field labels to task custom fields:
+A form can start an automation that creates tasks. Taskade then matches the form field labels to the task custom fields:
-* **Automatic Matching**: "Priority Level" dropdown → Priority custom field
-* **Type Recognition**: Date fields map to date custom fields automatically
-* **Dynamic Options**: Dropdown values sync with custom field options
+* **Name match**: A "Priority Level" dropdown goes to the Priority custom field
+* **Type match**: A date field goes to a date custom field
+* **Dropdown options**: The dropdown values match the options of the custom field
-### **Professional Use Cases**
+### **Example Use Cases**
**Document Approval Workflow:**
@@ -1228,7 +1237,7 @@ Document Collection → Verification → CRM Update → Welcome Email
### **File Type Support**
-Automations now handle a wide range of file types:
+Automations can process these file types:
* **Documents**: PDF, DOC, DOCX, TXT, RTF
* **Images**: JPG, PNG, GIF, WebP, SVG
diff --git a/genesis-living-system-builder/automation/advanced-actions.md b/genesis-living-system-builder/automation/advanced-actions.md
index b5d7c3c..40995e7 100644
--- a/genesis-living-system-builder/automation/advanced-actions.md
+++ b/genesis-living-system-builder/automation/advanced-actions.md
@@ -5,10 +5,10 @@ description: >-
# Advanced Automation Actions
-**Master the latest automation capabilities including loop processing, WhatsApp Business integration, HTTP response schema generation, and error handling. These features turn complex workflows into simple, reliable automations.**
+This page shows how to repeat steps for each item in a list and send WhatsApp Business messages. It also shows how to read API responses and control what happens when a step fails.
{% hint style="info" %}
-The flows on this page use the automation editor. You add each step from the step picker, and you insert values from earlier steps with the `@` picker. Premium integrations (for example WhatsApp Business, Slack, X/Twitter, LinkedIn and Facebook Pages) need a paid plan.
+The flows on this page use the automation editor. You add each step from the step picker. You insert values from earlier steps with the `@` picker. Premium integrations need a paid plan, for example WhatsApp Business, Slack, X/Twitter, LinkedIn, and Facebook Pages.
{% endhint %}
---
@@ -17,17 +17,17 @@ The flows on this page use the automation editor. You add each step from the ste
### **What Are Loop Actions?**
-A Loop repeats a set of steps once for each item in a list. One trigger fans out into many personalized actions, one per row, task, or item. Loops support filters, conditional branches, and nested loops.
+A Loop repeats a set of steps one time for each item in a list. One trigger can start one personalized action for each row, task, or item. A loop can contain filters, branches, and other loops.
#### **For Each Loop**
-A loop is a bounded for-each. It runs once for every item that already exists in the list, then stops. It does not repeat until a goal is reached, and it has no exit condition of its own.
+A loop runs one time for each item that is in the list when the loop starts. Then it stops. It does not repeat until a goal occurs, and it has no exit condition of its own.
-Point the loop at a list in one of two ways:
+Under **Variable**, select the source of the list:
-| Field | Description |
+| Option | Description |
| --- | --- |
-| **List** | Type a custom list of items by hand. The loop processes each item in order. |
-| **Variable** | Connect the loop to a list from an earlier step, such as **Find Row(s)** in Google Sheets or **Find Task(s)** in Taskade. |
+| **Manual List** | You type a custom list of items. The loop uses the items in their order. |
+| **Link to Variable** | You connect the loop to a list from an earlier step, for example **Find Row(s)** in Google Sheets or **Find Task(s)** in Taskade. |
```
Trigger: Schedule (weekly)
@@ -40,7 +40,7 @@ Loop: For each row
### **Advanced Loop Features**
#### **Nested Loops**
-Put a loop inside another loop when each item holds its own list, for example projects that each hold tasks.
+If each item holds its own list, put a loop inside another loop. For example, each project can hold tasks.
```
Action: Find Task(s) (tasks to review)
@@ -50,14 +50,14 @@ Loop: For each task
```
#### **Skip Items in a Loop**
-A loop has no break condition. To skip items, add a **Filter Data** step inside the loop. An item that does not match stops there, and the loop continues with the next item.
+A loop has no break condition. To skip items, add a **Filter Data** step inside the loop. An item that does not match stops at the filter. Then the loop continues with the next item.
-| What happens | Result |
+| What occurs | Result |
| --- | --- |
-| The list is empty | The loop runs zero times, and the automation continues after it. |
-| A step inside the loop fails | The loop stops at that item and the run is recorded as **Failed**. Earlier items keep their work. |
+| The list is empty | The loop runs zero times. The automation continues after the loop. |
+| A step inside the loop fails | The loop stops at that item, and the run shows as **Failed**. Earlier items keep their work. |
| A failed step has **Continue if this step fails** checked, with **Stop here** | The loop moves to the next item. |
-| A failed step has **Continue if this step fails** checked, with **Go to the next step** | The next step inside the loop runs for the same item and reads the result of the failed step as empty. |
+| A failed step has **Continue if this step fails** checked, with **Go to the next step** | The next step inside the loop runs for the same item. It reads the result of the failed step as empty. |
---
@@ -66,25 +66,28 @@ A loop has no break condition. To skip items, add a **Filter Data** step inside
### **Setup & Authentication**
#### **WhatsApp Business API Configuration**
-1. Add a WhatsApp Business action to an automation and click **Connect**.
-2. In [Meta for Developers](https://developers.facebook.com/apps/), create a **Business** app and click **Set Up WhatsApp**.
-3. Link your Business Portfolio and click **Start using API**.
-4. Click **Generate Access Token**. Paste it in the **System User Access Token** field.
-5. Copy the **WhatsApp Business Account ID** from the same page.
+1. Add a WhatsApp Business action to an automation.
+2. Click **Connect**.
+3. In [Meta for Developers](https://developers.facebook.com/apps/), create a **Business** app.
+4. Click **Set Up WhatsApp**.
+5. Link your Business Portfolio, then click **Start using API**.
+6. Click **Generate Access Token**.
+7. Paste the token in the **System User Access Token** field.
+8. Copy the **WhatsApp Business Account ID** from the same page.
{% hint style="info" %}
-**Building your own webhook receiver?** See the developer [Webhooks reference](../../apis-living-system-development/webhooks.md) for payload formats and event types.
+**Do you build your own webhook receiver?** The developer [Webhooks reference](../../apis-living-system-development/webhooks.md) gives the payload formats and the event types.
{% endhint %}
### **Message Types & Templates**
-A trigger from any source starts the automation, and the WhatsApp step sends the message. WhatsApp Business also has a New Message Received trigger.
+A trigger from any source starts the automation. Then the WhatsApp step sends the message. WhatsApp Business also has a New Message Received trigger.
| Action | Description |
| --- | --- |
-| **Send Message** | Sends a text message to a contact or group |
-| **Send Template Message** | Sends a pre-approved template message |
-| **Send Media** | Sends an image or other media file |
+| **Send Message** | Sends a text message to a contact or a group |
+| **Send Template Message** | Sends a template message that Meta approved |
+| **Send Media** | Sends an image or another media file |
#### **Text Messages**
```
@@ -129,20 +132,28 @@ Action: WhatsApp Business → Send Message
Text: "Your order is being prepared!"
```
+#### **If a message does not arrive**
+
+| What you see | What to do |
+| --- | --- |
+| "WhatsApp delivers a free-form message only within 24 hours" (Meta error 131047) | Use **Send Template Message**. |
+| "WhatsApp could not deliver to this number" (Meta error 131026) | Make sure that the number is correct and includes the country code. If the person does not use WhatsApp, contact them another way. |
+| "The WhatsApp access token expired or was revoked" (Meta error 190) | Connect again with a permanent System User token. |
+
---
## HTTP Schema Generator
### **Auto-Generate API Schemas**
-The **Send HTTP Request** action can build a response schema for you. Later steps can then pick fields from the response with the `@` picker.
+The **Send HTTP Request** action can make a response schema for you. A schema lists the fields of the response. Later steps can then select these fields with the `@` picker.
#### **Automatic Schema Detection**
-1. Set up the HTTP request as usual.
+1. Configure the HTTP request.
2. Under **Response Body Schema**, select **JSON**.
-3. Click **Generate Automatically**. A dialog opens where you send a test request.
-4. Taskade extracts the available fields from the JSON response.
-5. Review the extracted fields, and keep or remove each one.
+3. Click **Generate Automatically**.
+4. In the dialog, send a test request. Taskade reads the fields from the JSON response.
+5. Examine the fields. Keep or remove each field.
#### **Generated Schema Output**
```json
@@ -157,7 +168,7 @@ The **Send HTTP Request** action can build a response schema for you. Later step
```
### **Schema-Driven Validations**
-To check a response before you act on it, add a **Branch** or **Filter Data** step after the request. For example, continue only when `email` is not empty. To stop the run on bad data, use **Stop and mark as failed**.
+To examine a response before the next steps use it, add a **Branch** or **Filter Data** step after the request. For example, continue only when `email` is not empty. To stop the run when the data is bad, use **Stop and mark as failed**.
---
@@ -174,7 +185,7 @@ To check a response before you act on it, add a **Branch** or **Filter Data** st
```
Trigger: Task Completed (launch checklist)
-Action: Generate with AI (write a short launch post)
+Action: Write with AI (write a short launch post)
Action: X/Twitter → Create Post
```
@@ -198,9 +209,9 @@ Action: X/Twitter → Create Post (summary + link)
#### **Multi-Format Content Publishing**
| Action | Description |
| --- | --- |
-| **Create Post** | Creates a post on a Facebook Page you manage |
-| **Create Post with Image** | Creates a post with an image on a Facebook Page you manage |
-| **Create Video Post** | Creates a video post on a Facebook Page you manage |
+| **Create Post** | Creates a post on a Facebook Page that you manage |
+| **Create Post with Image** | Creates a post with an image on a Facebook Page that you manage |
+| **Create Video Post** | Creates a video post on a Facebook Page that you manage |
---
@@ -209,22 +220,30 @@ Action: X/Twitter → Create Post (summary + link)
### **Smart Error Recovery**
#### **Automatic Retry Strategies**
-Taskade retries a step automatically after a temporary error, such as a network error, an API rate limit, or a server error. An agent step (**Ask an agent**, **Ask an agent team**, or **Run an agent command**) runs once and is not retried.
+Taskade tries a step again automatically after a temporary error. Examples are a network error, an API rate limit, or a server error. An agent step runs one time only, and Taskade does not try it again. The agent steps are **Ask an agent**, **Ask an agent team**, and **Run an agent command**.
+
+You can select what occurs when a step still fails. Open an action step and check **Continue if this step fails**. Then select one option:
-To decide what happens when a step still fails, open an action step and check **Continue if this step fails**. Then pick one option:
+* **Stop here** ends the work on that item. In a loop, the loop moves to the next item. Outside a loop, the automation stops at that step, and the run does not show as failed.
+* **Go to the next step** runs the next step as if the failed step returned nothing. The run does not show as failed.
-* **Stop here** ends the work on that item. In a loop, the loop moves to the next item. Outside a loop, the automation stops at that step, and the run is not marked as failed.
-* **Go to the next step** runs the next step as if the failed step returned nothing. The run is not marked as failed.
+Some steps also have an **Error Behavior** setting with **Continue** or **Fail**. Read a web page is one of these steps.
-Some steps, such as Scrape Webpage, also have an **Error Behavior** setting with **Continue** or **Fail**.
+#### **Retry a step when an app is down**
+Open an action step and check **Retry if this step fails**. Select how long to **Wait** (1 to 5 minutes). Then select how many times to try again (1 to 5).
+
+This option helps when a connected app is busy or down for a few minutes. A bad input or a broken connection still fails at once. Retries can delay later runs of the same automation. A step that sends a message can send it two times. An agent step does not have this option.
+
+#### **Pause for a person**
+Add a **Wait for approval** step. The run waits until a teammate approves or rejects it. A rejection stops the run.
#### **Context-Aware Error Classification**
-Use a **Branch** step after a risky step to route on its result. Use **Stop and mark as failed** to end a run on purpose when the data is wrong.
+To send the flow on a different path after a step that can fail, add a **Branch** step after it. To end a run on purpose when the data is wrong, use **Stop and mark as failed**.
### **Comprehensive Error Monitoring**
#### **Run history**
-The automation **Runs** tab lists every run. Open a run to see the exact input and output of each step and the reason a step failed or was skipped. Click **↻** to retry a failed run after you fix the cause.
+The automation **Runs** tab lists every run. Open a run to see the input and the output of each step. The run also shows why a step failed or why Taskade skipped it. After you correct the cause, click **↻** to run a failed run again.
---
@@ -235,27 +254,27 @@ The automation **Runs** tab lists every run. Open a run to see the exact input a
#### **Dynamic Agent Selection**
| Action | Description |
| --- | --- |
-| **Ask an agent** | Uses a custom AI agent to respond to a query |
-| **Run an agent command** | Runs a custom agent command |
-| **Ask an agent**, with **Output Type** set to **Structured Output** | Returns typed fields you can use in later steps |
+| **Ask an agent** | Uses a custom AI agent to answer a query |
+| **Run an agent command** | Runs a command of a custom agent |
+| **Ask an agent**, with **Output Type** set to **Structured Output** | Returns typed fields that you can use in later steps |
| **Create an agent** | Creates a new AI agent in the workspace |
| **Update an agent** | Changes the name, the greeting, or the instructions of an agent |
-To pick an agent per case, add a **Branch** step and call a different agent on each path.
+To use a different agent for each case, add a **Branch** step. Then call a different agent on each path.
#### **Agent Collaboration Workflows**
-The **Ask an agent team** action sends one prompt to an AI Team and returns responses from multiple agents. AI Teams need Pro and up. The **Agent Tool** trigger lets an agent run an automation as one of its tools during a conversation.
+The **Ask an agent team** action sends one prompt to an AI Team. It returns the answers of more than one agent. AI Teams need Pro and up. With the **Agent Tool** trigger, an agent can run an automation as one of its tools during a chat.
### **Conversation Export & Analysis**
#### **Automated Conversation Processing**
-The **Agent Public Chat Ended** trigger starts an automation when a public agent conversation stays inactive for 5 minutes. The trigger passes the full transcript to later steps.
+The **Agent Public Chat Ended** trigger starts an automation when a public agent chat has no activity for 5 minutes. The trigger sends the full transcript to later steps.
```
Trigger: Agent Public Chat Ended
Action: Categorize with AI (Sentiment Analysis)
Action: Google Sheets → Insert Row (transcript + sentiment)
-Action: Add Knowledge to Agent (transcript)
+Action: Add knowledge to an agent (transcript)
```
---
@@ -265,30 +284,28 @@ Action: Add Knowledge to Agent (transcript)
### **Batch Processing Optimization**
#### **Intelligent Batching**
-* Put a **Filter Data** step early, so later steps run only for the items you need.
-* Keep expensive steps, such as AI steps, after the filter inside a loop.
-* Use **Delay** to space out calls to a service with rate limits.
+* Put a **Filter Data** step early. Then later steps run only for the items that you need.
+* Inside a loop, put steps that cost more, such as AI steps, after the filter.
+* If a service has rate limits, add **Delay** steps between the calls to it.
### **Resource Management**
#### **Memory & Processing Limits**
-Automation runs count toward the plan limit for runs per billing period. AI steps use AI credits, and utility actions use no AI credits. See [pricing](https://www.taskade.com/pricing) for the limits of each plan.
+Each automation run counts toward the limit of runs for each billing period of your plan. AI steps use AI credits. Utility actions use no AI credits. The [pricing](https://www.taskade.com/pricing) page gives the limits of each plan.
---
## Build automations with AI
-You don't have to assemble flows by hand — Taskade already builds automations from natural language:
+You do not have to build each flow step by step. Taskade can make an automation from a plain-English description:
-* **Taskade EVE** — describe the outcome you want and Taskade EVE drafts the automation: trigger, actions, and connected integrations.
-* **Workflow Generator** at [taskade.com/create/flow](https://www.taskade.com/create/flow), or **Add automation → Generate With AI** in the Automations tab — describe the workflow and pick a mode: Auto, Agent Tool, Workflow, or Form.
-* Generated flows are normal automations — open them in the flow editor to adjust triggers, actions, and conditions like any hand-built workflow.
+* **Taskade EVE**: Describe the result that you want. Taskade EVE writes the automation with its trigger, actions, and connected integrations.
+* **Workflow Generator**: Go to [taskade.com/create/flow](https://www.taskade.com/create/flow), or click **Add automation → Generate With AI** in the Automations tab. Describe the workflow and select a mode: Auto, Agent Tool, Workflow, or Form.
+* **Generated flows**: A generated flow is a normal automation. You can open it in the flow editor and change its triggers, actions, and conditions.
---
-**🚀 Ready to use these advanced automation capabilities?** Start with loop actions to handle batch processing, then explore WhatsApp Business integration to reach customers where they are.
-
-*For more automation examples and tutorials, check out our [Automation Recipes](./recipes.md) and [Integration Guide](./integrations.md).*
+For more automation examples, read [Automation Recipes](./recipes.md) and the [Integration Guide](./integrations.md).
---
diff --git a/genesis-living-system-builder/automation/ai-forms.md b/genesis-living-system-builder/automation/ai-forms.md
index 6bce2aa..d94612a 100644
--- a/genesis-living-system-builder/automation/ai-forms.md
+++ b/genesis-living-system-builder/automation/ai-forms.md
@@ -1,14 +1,14 @@
---
description: >-
- Build Taskade AI Forms that start an automation on every submission, then use AI steps to analyze, categorize, and route responses into your workspace and CRM.
+ Build Taskade AI Forms that start an automation on each submission. AI steps then analyze, sort, and route the responses to your workspace and CRM.
---
# AI Forms: Intelligent Data Collection
-**Create forms that start an automation the moment someone submits them. AI Forms collect the data, and the AI steps in your automation analyze each response, categorize it, and run the follow-up actions.**
+With AI Forms, you can collect answers and act on them at once. Each submission starts an automation. The AI steps in that automation read each answer, sort it, and run the follow-up actions.
{% hint style="success" %}
-**New to AI Forms?** Start with a simple feedback form, then add one AI step after the trigger to see how a basic form turns into an automated workflow.
+**Is this your first AI Form?** Start with a simple feedback form. Then add one AI step after the trigger and see the result of each submission.
{% endhint %}
***
@@ -19,19 +19,19 @@ description: >-
| Feature | Traditional Forms | AI Forms |
| --------------------- | ----------------------------- | ------------------------------------------------- |
-| **Data Processing** | Store raw responses | AI steps analyze each response |
-| **Follow-up Actions** | Manual review required | Automation routes each submission |
-| **Integration** | Basic webhook/email | Taskade projects, agents, and 100+ integrations |
-| **Result Display** | Static thank-you page | Custom end page after submission |
+| **Data Processing** | Keep the raw answers | AI steps read and sort each answer |
+| **Follow-up Actions** | A person reviews each answer | The automation sends each submission to the correct place |
+| **Integration** | A basic webhook or email | Taskade projects, agents, and 100+ integrations |
+| **Result Display** | A fixed thank-you page | A custom end page after the submission |
### **Key AI Capabilities**
-The form is the trigger. The AI work happens in the steps that follow it:
+The form is the trigger. The AI work occurs in the steps after the trigger:
-* **Sentiment Analysis**: Add **Categorize with AI** with the Sentiment Analysis preset to detect the tone of text responses
-* **Auto-categorization**: Use **Categorize with AI** or **Ask Agent With Structured Output** to tag and organize responses
-* **Intelligent Routing**: Add a **Branch** step on the AI result to pick the next action
-* **File Reading**: Add **Convert File to Text** to read an uploaded file, including a scanned PDF or a phone photo
+* **Sentiment Analysis**: Add **Categorize with AI** with the Sentiment Analysis preset. It finds the tone of text answers.
+* **Auto-categorization**: Use **Categorize with AI**, or an **Ask an agent** step with structured output. These steps tag and sort the answers.
+* **Intelligent Routing**: Add a **Branch** step on the AI result. The branch selects the next action.
+* **File Reading**: Add **Turn a file into text** to read an uploaded file. It also reads a scanned PDF or a photo from a phone.
***
@@ -41,7 +41,9 @@ The form is the trigger. The AI work happens in the steps that follow it:
#### **Option A: AI Generation (Recommended)**
-In the Automations tab, click **Add automation → Generate With AI** and pick **Form** mode. Describe the form and what happens after a submission:
+1. In the Automations tab, click **Add automation → Generate With AI**.
+2. Select **Form** mode.
+3. Describe the form and the actions after a submission:
```
"Create a customer satisfaction survey for my restaurant with:
@@ -54,42 +56,42 @@ In the Automations tab, click **Add automation → Generate With AI** and pick *
#### **Option B: Form Builder Interface**
-1. Go to your workspace **Automations** tab
-2. Click **Add automation**, then pick a template, or click **Create** to start from scratch
-3. Click **➕ Add Trigger → Form**
-4. Configure form settings and fields in the sidebar
-5. Add automation actions for processing responses
+1. Open the **Automations** tab of your workspace.
+2. Click **Add automation**. Then select a template, or click **Create** to start with an empty automation.
+3. Click **➕ Add Trigger → Form**.
+4. Configure the form settings and the fields in the sidebar.
+5. Add the automation actions that process the answers.
### **Step 2: Configure Form Settings**
| Setting | Description | Example |
| ---------------------- | ---------------------------------------- | ------------------------------------------ |
-| **Form Title** | Choose a unique title for your form | "Customer Feedback Survey" |
-| **Form Description** | Describe the purpose of the form | "Help us improve your experience" |
-| **Submit Button Text** | Choose a title for the submission button | "Submit Feedback" |
-| **Form End Page** | Display an end screen with custom text | "Thank you! We'll respond within 24 hours" |
+| **Form Title** | A unique title for your form | "Customer Feedback Survey" |
+| **Form Description** | The purpose of the form | "Help us improve your experience" |
+| **Submit Button Text** | The text on the submit button | "Submit Feedback" |
+| **Form End Page** | An end screen with custom text | "Thank you! We'll respond within 24 hours" |
### **Step 3: Choose Form Fields**
-AI Forms include customizable field types to collect different kinds of information:
+AI Forms have many field types. Each type collects a different kind of information:
-| Field Type | Description | Best For | Pair With |
+| Field Type | Description | Use It For | Next Step to Add |
| ------------------- | --------------------------------------- | ------------------------------ | ------------------------------------------ |
-| **Short Text** | For short answers like a name or title | Names, titles, keywords | Add Task, Insert Row |
-| **Long Text** | For longer answers like comments | Feedback, descriptions | Categorize with AI (sentiment, topic) |
-| **Option Dropdown** | A list of choices where one is selected | Categories, preferences | Branch on the selected option |
-| **Number** | For numbers like age or quantity | Ages, quantities, scores | Filter Data on a range |
-| **File** | Lets users upload a file | Documents, images, attachments | Convert File to Text, then Ask Agent |
-| **Phone** | For phone numbers | Contact information | Twilio Send SMS, WhatsApp Business |
-| **Email** | For email addresses | Contact information | Gmail Send Email, CRM contact |
-| **Date** | A calendar to pick a date | Appointments, deadlines | Google Calendar Create Event |
-| **Rating** | For giving a rating | Satisfaction, quality scores | Branch on low ratings |
-| **Single-Select** | Allow one option from predefined list | Multiple choice questions | Branch on the answer |
-| **Multi-Select** | Allow multiple options | Interests, features, tags | Loop over the selected options |
-| **Checkbox** | A yes/no checkbox submitting true/false | Consent boxes, opt-ins, toggles | Branch or filter automations on the value |
-| **URL** | Accept web addresses or links | Website links, references | Scrape Webpage |
-
-A **File** upload is saved with the submission and shows up in the Media tab.
+| **Short Text** | A short answer, like a name or a title | Names, titles, keywords | Add Task, Insert Row |
+| **Long Text** | A longer answer, like a comment | Feedback, descriptions | Categorize with AI (sentiment, topic) |
+| **Option Dropdown** | A list of choices. The user selects one | Categories, preferences | Branch on the selected option |
+| **Number** | A number, like an age or a quantity | Ages, quantities, scores | Filter Data on a range |
+| **File** | The user uploads a file | Documents, images, attachments | Turn a file into text, then Ask an agent |
+| **Phone** | A phone number | Contact information | Twilio Send SMS, WhatsApp Business |
+| **Email** | An email address | Contact information | Gmail Send Email, CRM contact |
+| **Date** | A calendar to select a date | Appointments, deadlines | Google Calendar Create Event |
+| **Rating** | A rating | Satisfaction, quality scores | Branch on low ratings |
+| **Single-Select** | One option from a list | Multiple-choice questions | Branch on the answer |
+| **Multi-Select** | More than one option from a list | Interests, features, tags | Loop over the selected options |
+| **Checkbox** | A yes or no box. It sends true or false | Consent boxes, opt-ins, toggles | Branch or Filter Data on the value |
+| **URL** | A web address or a link | Website links, references | Read a web page |
+
+Taskade saves a **File** upload with the submission. The file shows in the Media tab.
### **Step 4: Configure AI-Powered Processing**
@@ -115,7 +117,7 @@ Action: Update Custom Fields (Sentiment)
#### **Dynamic Question Logic**
-Form questions are fixed. To act differently on different answers, add **Branch** steps after the trigger.
+The questions of a form do not change. To do different actions for different answers, add **Branch** steps after the trigger.
### **Step 5: Set Up AI Processing Rules**
@@ -123,7 +125,7 @@ Form questions are fixed. To act differently on different answers, add **Branch*
```
Trigger: Form
-Action: Ask Agent With Structured Output
+Action: Ask an agent with structured output
Output: Category (Single Select:
complaint, compliment, suggestion, question)
Urgent (Boolean)
@@ -132,7 +134,7 @@ Branch: If Urgent is true → Slack → Send Direct Message (on-call)
#### **Smart Validation Rules**
-Each field type checks its own format, for example Email, Phone, Number, Date, and URL. For checks after submission, add a **Filter Data** step, or an **Ask Agent With Structured Output** step that flags incomplete answers.
+Each field type examines its own format, for example Email, Phone, Number, Date, and URL. To examine answers after a submission, add a **Filter Data** step. You can also add an **Ask an agent** step with structured output that finds incomplete answers.
***
@@ -146,14 +148,14 @@ Set the title, the description, the submit button text, and the end page text in
#### **Dynamic Visual Elements**
-* **Hide Branding**: Remove the "Powered by Taskade" footer from your public form (paid plans)
-* **Form End Page**: Show a custom end screen after submission
+* **Hide Branding**: Removes the "Powered by Taskade" footer from your public form (paid plans).
+* **Form End Page**: Shows a custom end screen after a submission.
### **Multi-Step Forms with AI Logic**
#### **Smart Step Progression**
-A form is one page. To collect more information later, send a second form link from the automation, for example in a follow-up email.
+A form has one page only. To collect more information later, send a link to a second form from the automation. For example, put the link in a follow-up email.
***
@@ -161,38 +163,45 @@ A form is one page. To collect more information later, send a second form link f
### **Start with AI Form Templates**
-Don't want to build forms from scratch? Click **➕ Add Trigger**, go to the **Forms** category, and choose a template on the right. Then customize it. Common uses:
+You can start from a template instead of an empty form:
+
+1. Click **➕ Add Trigger**.
+2. Open the **Forms** category.
+3. Select a template on the right.
+4. Change the template to fit your needs.
+
+Common templates:
| Template Type | Use Case | Included Fields | Automation Actions |
| ---------------------- | -------------------------------- | --------------------------------------- | ---------------------------------------- |
-| **Customer Feedback** | Collect product/service feedback | Rating, comments, contact info | Auto-categorize, notify team |
-| **Lead Qualification** | Qualify potential customers | Company, budget, timeline | Score leads, assign to sales |
-| **Support Ticket** | Technical support requests | Issue type, priority, description | Route to specialists, create tasks |
-| **Event Registration** | Event sign-ups and RSVPs | Name, email, dietary restrictions | Send confirmations, update attendee list |
+| **Customer Feedback** | Collect feedback on a product or service | Rating, comments, contact info | Sort the feedback, notify the team |
+| **Lead Qualification** | Qualify possible customers | Company, budget, timeline | Score leads, assign them to sales |
+| **Support Ticket** | Technical support requests | Issue type, priority, description | Send to specialists, create tasks |
+| **Event Registration** | Event sign-ups and RSVPs | Name, email, dietary restrictions | Send confirmations, update the attendee list |
| **Job Application** | Collect job applications | Resume upload, experience, availability | Screen candidates, schedule interviews |
-| **Survey Research** | Market research and surveys | Multiple choice, ratings, demographics | Analyze trends, generate reports |
+| **Survey Research** | Market research and surveys | Multiple choice, ratings, demographics | Find trends, make reports |
### **Share and Use Forms**
-1. Go to the **Automations** tab and open your form automation.
-2. Go to the **Form** tab at the top.
-3. Use the run button to test the form.
-4. Click **Publish** in the top-right corner to share the form.
+1. Open the **Automations** tab, then open your form automation.
+2. Open the **Form** tab at the top.
+3. Click the run button to test the form.
+4. To share the form, click **Publish** in the top-right corner.
#### **Public Link Sharing**
-* Share the form link by email, social media, or messaging
-* Publish forms on your own domain with a [custom domain](https://www.taskade.com/learn/genesis/custom-domains) (Pro and up)
-* See every submission in the automation **Runs** tab
+* You can share the form link by email, on social media, or in a message.
+* You can publish forms on your own domain with a [custom domain](https://www.taskade.com/learn/genesis/custom-domains) (Pro and up).
+* The automation **Runs** tab shows every submission.
#### **Website Embedding**
-You can embed a published form on your website as well as share its link. Click **Publish**, then copy the **Embed Code**.
+You can also put a published form on your website. Click **Publish**, then copy the **Embed Code**.
#### **Configure Form Visibility**
-* **Public link**: Anyone with the link can open and submit the form
-* **Require Password**: Anyone who opens the link must enter the password before viewing or submitting (Pro and up)
+* **Public link**: Anyone with the link can open and submit the form.
+* **Require Password**: A visitor must enter the password before the visitor can see or submit the form (Pro and up).
***
@@ -211,10 +220,10 @@ Action: Add Task (customer name, rating, and category)
#### **AI Agent Collaboration**
-* **Response Analysis**: An **Ask Agent** step reviews each submission
-* **Quality Scoring**: **Ask Agent With Structured Output** returns a score field
-* **Follow-up Suggestions**: The agent drafts the next step, and a person reviews it
-* **Pattern Recognition**: A weekly **Schedule** automation with **Find Task(s)** and **Ask Agent** summarizes trends
+* **Response Analysis**: An **Ask an agent** step reviews each submission.
+* **Quality Scoring**: An **Ask an agent** step with structured output returns a score field.
+* **Follow-up Suggestions**: The agent writes a draft of the next step. Then a person reviews it.
+* **Pattern Recognition**: A weekly **Schedule** automation with **Find Task(s)** and **Ask an agent** summarizes trends.
### **External Service Integration**
@@ -253,18 +262,18 @@ Branch on the Lead Qualification result:
#### **Completion Intelligence**
-The automation **Runs** tab lists each submission with the input and output of every step.
+The automation **Runs** tab lists each submission. It shows the input and the output of each step.
#### **Response Quality Metrics**
-To track quality, save the AI results of each submission (for example sentiment and category) to a project or a Google Sheet. Then review or chart that data.
+To track quality, save the AI results of each submission to a project or a Google Sheet. Examples are the sentiment and the category. Then review that data or make a chart of it.
### **Smart Reporting Dashboard**
#### **Auto-Generated Insights**
-* **Weekly Intelligence Reports**: A **Schedule** automation finds the week's submissions and asks an agent to summarize trends
-* **Actionable Recommendations**: Ask the agent to suggest form and process improvements in the same summary
+* **Weekly Intelligence Reports**: A **Schedule** automation finds the submissions of the week. Then it asks an agent to summarize trends.
+* **Actionable Recommendations**: In the same summary, the agent can suggest improvements to the form and to the process.
***
@@ -274,13 +283,19 @@ To track quality, save the AI results of each submission (for example sentiment
#### **Direct Website Embed**
-Publish the form, then share its link or embed it on your website.
+Publish the form. Then share its link or put it on your website.
### **API Access**
#### **Form Data**
-AI Forms does not currently expose a public REST API for reading submissions in API v1 (there is no `/forms` endpoint). To work with submission data programmatically, capture it with an **automation** when a form is submitted — route responses into a project, forward them to an external service via the HTTP/Webhook action, or hand them to an agent. See [Webhooks](../../apis-living-system-development/webhooks.md) and the [Action & Trigger Reference](./actions.md).
+API v1 has no public REST endpoint to read form submissions (there is no `/forms` endpoint). To use submission data in your own code, catch it with an **automation** when a user submits the form. The automation can:
+
+* Add the answers to a project
+* Send the answers to an outside service with the HTTP or Webhook action
+* Give the answers to an agent
+
+Read [Webhooks](../../apis-living-system-development/webhooks.md) and the [Action & Trigger Reference](./actions.md).
### **Webhook Integration**
@@ -303,30 +318,30 @@ Action: Send HTTP Request
#### **AI-Driven UX Principles**
-* **Start small**: Ask only for the fields your automation uses
-* **Use the right field type**: Email, Phone, Date, and URL fields check their own format
-* **Use a Checkbox for consent**: Branch or filter on the true/false value
+* **Start small**: Ask only for the fields that your automation uses.
+* **Use the correct field type**: Email, Phone, Date, and URL fields examine their own format.
+* **Use a Checkbox for consent**: Branch or filter on the true or false value.
#### **Response Quality Enhancement**
-* **Question Clarity**: Write short, specific questions
-* **Response Validation**: Add a Filter Data step, or an AI step that flags incomplete answers
-* **Test first**: Submit a test entry and read the run before you publish
+* **Question Clarity**: Write short and specific questions.
+* **Response Validation**: Add a Filter Data step, or an AI step that finds incomplete answers.
+* **Test first**: Before you publish, submit a test entry and read the run.
### **Data Privacy & Security**
#### **AI-Enhanced Privacy Controls**
-* **Require Password**: Protect a form that is not for the public (Pro and up)
-* **Collect less**: Do not ask for data the automation does not need
-* **Review the steps**: Check which connected tools receive each submission
+* **Require Password**: Protect a form that is not for the public (Pro and up).
+* **Collect less**: Do not ask for data that the automation does not need.
+* **Review the steps**: Find out which connected tools get each submission.
### **Performance Optimization**
#### **Loading Speed & Responsiveness**
-* Put **Filter Data** early, so later steps run only for submissions that need them
-* Keep AI steps after the filter to save AI credits
+* Put **Filter Data** early. Then later steps run only for the submissions that need them.
+* Put AI steps after the filter. This uses fewer AI credits.
***
@@ -337,7 +352,7 @@ Action: Send HTTP Request
```
Use Case: Restaurant feedback collection
Steps:
- - Convert File to Text or Ask Agent on uploaded meal photos
+ - Turn a file into text or Ask an agent on uploaded meal photos
- Categorize with AI (Sentiment Analysis) on written feedback
- Branch: route complaints to managers
- Weekly Schedule automation: summarize service quality trends
@@ -358,7 +373,7 @@ Steps:
```
Use Case: Conference attendee registration
Steps:
- - Ask Agent: recommend sessions from the attendee's interests
+ - Ask an agent: recommend sessions from the attendee's interests
- Gmail → Send Email (confirmation with the recommendations)
- Google Sheets → Insert Row (attendee list)
```
@@ -371,29 +386,27 @@ Steps:
#### **AI Analysis Not Working**
-* **Check AI credits**: AI steps use the workspace AI credits
-* **Check the runs**: Open the run in the **Runs** tab and read the input and output of each step
-* **Response Quality**: AI steps need meaningful text responses for analysis
-* **Check the input**: Make sure that the AI step points at the right form field with the `@` picker
+* **AI credits**: AI steps use the AI credits of the workspace. Make sure that credits are available.
+* **Runs**: Open the run in the **Runs** tab. Read the input and the output of each step.
+* **Response Quality**: An AI step needs a text answer with real content.
+* **Input**: Make sure that the AI step points at the correct form field. Use the `@` picker.
#### **Integration Problems**
-* **Plan**: Premium integrations need a paid plan
-* **Data Mapping**: Verify the field mapping between the form and the destination step
-* **Authentication**: Reconnect the integration if its connection expired
-* **Rate Limits**: Add a Delay step for a service with rate limits
+* **Plan**: Premium integrations need a paid plan.
+* **Data Mapping**: Make sure that each form field goes to the correct field of the next step.
+* **Authentication**: If the connection of an integration expired, connect it again.
+* **Rate Limits**: If a service has rate limits, add a Delay step.
### **Getting Help**
* **Documentation**: [Taskade Developer API](../../apis-living-system-development/developer-home.md)
-* **Support**: Contact support for advanced configuration assistance
-* **Templates**: Browse pre-built AI form templates in your workspace
+* **Support**: Contact support for help with advanced configuration.
+* **Templates**: Find ready-made AI form templates in your workspace.
***
-**🚀 Ready to transform your data collection with AI?** Start with a simple feedback form and discover how AI steps can make every form submission more valuable and actionable.
-
-_For more advanced AI integrations, explore our_ [_AI Agents documentation_](../ai-features/ai-agents-getting-started.md) _and_ [_Automation workflows_](./)_._
+_For more AI integrations, read the_ [_AI Agents documentation_](../ai-features/ai-agents-getting-started.md) _and_ [_Automation workflows_](./)_._
---
diff --git a/genesis-living-system-builder/automation/comprehensive-integration-guide.md b/genesis-living-system-builder/automation/comprehensive-integration-guide.md
index 8b52f7a..4393272 100644
--- a/genesis-living-system-builder/automation/comprehensive-integration-guide.md
+++ b/genesis-living-system-builder/automation/comprehensive-integration-guide.md
@@ -1,40 +1,40 @@
---
description: >-
- Connect Taskade to Slack, Gmail, HubSpot, Sheets and more with trigger-and-action automations, plus custom HTTP, webhook, and API integrations.
+ Connect Taskade to Slack, Gmail, HubSpot, Google Sheets, and more with automations. Use HTTP requests, webhooks, or the API for other tools.
---
# Integration Overview
-**Transform isolated tools into intelligent, living ecosystems that learn, adapt, and evolve together.**
+**Connect Taskade to the tools that you use, and let your automations do the work between them.**
{% hint style="success" %}
-**Living Integrations** go beyond static connections—they create intelligent relationships between systems that learn from data flow patterns, adapt to changing business needs, and continuously optimize performance. Your connected tools become a living ecosystem that works smarter with every interaction.
+**Integrations** connect your workspace to other apps. An event in one app can start an automation. The automation can then write the result to another app or to a project.
{% endhint %}
## Integrations in Taskade
-Integrations connect Taskade to the tools you already use. They run through Automations using **triggers** (events) and **actions** (what happens next).
+Integrations connect Taskade to the tools that you use now. They run inside Automations. A **trigger** is the event that starts the automation. An **action** is the work that follows.
-If you want the bigger model, start with [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
+For the full model, start with [Workspace DNA](../../genesis-living-system-builder/genesis/workspace-dna.md).
### What you can do
-* **Notify and route work**: Slack/Teams alerts, assignments, escalations.
-* **Sync data**: push form submissions into a CRM, write rows to Sheets, create tickets.
-* **Move faster**: auto-create tasks, summaries, follow-ups, and reports.
-* **Close loops**: trigger workflows from app events and write results back into projects.
+* **Send and route work**: Send Slack or Microsoft Teams alerts, assign tasks, and escalate issues.
+* **Sync data**: Send form submissions to a CRM, write rows to Google Sheets, and create tickets.
+* **Save time**: Create tasks, summaries, follow-ups, and reports automatically.
+* **Close the loop**: Start a workflow from an app event, and write the result back into a project.
### Typical workflows
* Lead capture → qualify → create CRM record → notify team
-* Support email/webhook → categorize → create ticket → assign owner
-* Schedule trigger → generate report → send to stakeholders
+* Support email or webhook → categorize → create ticket → assign owner
+* Schedule trigger → write report → send to stakeholders
### Quick start
-1. Create an automation with a trigger + action.
-2. Add the tool's step and click **Connect** to link your account.
-3. Test with real data.
+1. Create an automation with a trigger and an action.
+2. Add the step for the tool. Click **Connect** to link your account.
+3. Test the automation with real data.
Start here:
@@ -44,45 +44,49 @@ Start here:
## Popular Integrations
-Most teams start with:
+Most teams start with these tools:
* Communication: Slack, Microsoft Teams, email
* Productivity: Google Workspace, Microsoft Outlook, Microsoft To Do
* Data: Google Sheets, Airtable
* CRM: HubSpot, Salesforce
-* Dev: GitHub, Jira
+* Development: GitHub, Jira
-Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot and Salesforce) need a paid plan. The Free plan runs the built-in steps and the standard connectors (HTTP Request, RSS, Pushover and MCP Client).
+Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, and Salesforce) need a paid plan. The Free plan runs the built-in steps and the standard connectors (HTTP Request, RSS, Pushover, and MCP Client).
-Full list and setup details: [Integration Options](integrations.md)
+For the full list and the setup details, read [Integration Options](integrations.md).
+
+## Files and Agent Knowledge
+
+You can also bring outside content into Taskade without an automation. Drop a file into a project, into the knowledge of an agent, or into **Media Files**. Taskade reads the text, and it can read the text in scans and images. Agent knowledge accepts every supported file format except images.
## Advanced Integration Patterns
-Start with one trigger and one outcome. Then add steps and branches.
+Start with one trigger and one result. Then add steps and branches.
Common patterns:
* Form submission → qualify → create CRM record → notify Slack
-* Support email/webhook → categorize → create ticket → assign owner
-* Schedule trigger → generate report → send digest
+* Support email or webhook → categorize → create ticket → assign owner
+* Schedule trigger → write report → send digest
-See the full building blocks:
+For all the building blocks, read:
* [Automation Getting Started](./)
* [Actions & Triggers](actions.md)
## Integration Best Practices
-Keep these four things tight:
+Follow these four rules:
-* **Scope**: grant the minimum permissions you need.
-* **Test**: run with real data, not placeholders.
-* **Handle failures**: decide what to do when an API call fails.
-* **Observe**: monitor errors and retries for critical workflows.
+* **Scope**: Give each connection only the permissions that it needs.
+* **Test**: Test with real data, not sample text.
+* **Failures**: Decide what the automation does when a step fails. You can check **Retry if this step fails** or **Continue if this step fails** on a step.
+* **Monitoring**: Examine the errors and retries of important workflows in the **Runs** tab.
## Custom Integration Development
-Don’t see your tool? Use HTTP requests/webhooks or the Taskade API.
+If your tool is not in the list, use an HTTP request, a webhook, or the Taskade API.
Start here:
@@ -91,24 +95,24 @@ Start here:
## Getting Started with Integrations
-### Planning Your Integration Strategy
+### Plan Your Integration
Keep it simple:
-1. Pick one workflow you want to automate end-to-end.
-2. Choose one trigger and one destination tool.
-3. Add steps only after the core path works.
+1. Select one workflow that you want to automate from start to end.
+2. Select one trigger and one destination tool.
+3. Add more steps only after the main path works.
### Implementation Checklist
* ✅ Connect the integration
* ✅ Build one automation
* ✅ Test with real data
-* ✅ Add notifications for failure cases
+* ✅ Add notifications for failed runs
## Support and Resources
-Use these when you get stuck:
+Use these pages when you need help:
* Full integration list: [Integration Options](integrations.md)
* Workflow basics: [Automation Getting Started](./)
diff --git a/genesis-living-system-builder/automation/comprehensive-integrations.md b/genesis-living-system-builder/automation/comprehensive-integrations.md
index 013dc28..a2a5de4 100644
--- a/genesis-living-system-builder/automation/comprehensive-integrations.md
+++ b/genesis-living-system-builder/automation/comprehensive-integrations.md
@@ -5,10 +5,12 @@ description: >-
# Comprehensive Integrations
-This document provides a reference of popular integrations, triggers, and actions in Taskade's automation system. It does not list every connector or every action. For the full, current list, open the step picker in the automation editor. For step-by-step guides, see the [Taskade Automation learn center](https://www.taskade.com/learn/automation). Every integration below also has a full setup page in the [Taskade Integrations directory](https://www.taskade.com/integrations).
+This page lists popular integrations with their triggers and actions. Each row gives the step ID, the step name, and its properties. The list does not show every connector or every action.
+
+To see the full, current list, open the step picker in the automation editor. For step-by-step guides, read the [Taskade Automation learn center](https://www.taskade.com/learn/automation). Each integration below also has a setup page in the [Taskade Integrations directory](https://www.taskade.com/integrations).
{% hint style="info" %}
-Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub and Stripe) need a paid plan. The Free plan runs the built-in Taskade steps and the standard connectors (HTTP Request, RSS, Pushover and MCP Client). The Every X Minutes schedule also needs a paid plan.
+Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub, and Stripe) need a paid plan. The Free plan runs the built-in Taskade steps and the standard connectors (HTTP Request, RSS, Pushover, and MCP Client). The Every X Minutes schedule also needs a paid plan.
{% endhint %}
## Core Taskade Tools
@@ -51,16 +53,16 @@ Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub a
| Action | Properties |
| --- | --- |
| `ai.ask` — Ask AI | `prompt`, `input` |
-| `ai.generate` — Generate with AI | `prompt`, `input` |
-| `ai.respond` — Respond with AI | `prompt`, `input` |
+| `ai.generate` — Write with AI | `prompt`, `input` |
+| `ai.respond` — Reply with AI | `prompt`, `input` |
| `ai.categorize` — Categorize with AI | `input`, `output` |
| `ai.prompt` — Prompt AI | `prompt`, `modelName` |
-| `agent.ask` — Ask Agent | `agentId`, `input`, `outputType`, `output` |
-| `agent.run_command` — Run Agent Command | `agentId`, `commandId`, `input` |
-| `agent.add_knowledge` — Add Knowledge To Agent | `agentId`, `knowledgeType`, `title`, `knowledge` |
-| `agent.add_knowledge.v2` — Add Knowledge To Agent (v2) | `agentId`, `knowledgeType`, `title`, `knowledge` |
-| `agent.add_project_as_knowledge` — Add Project to Agent Knowledge | `agentId`, `projectId` |
-| `agentTeam.ask` — Ask Agent Team | `agentTeamId`, `executionMode`, `input`, `output` |
+| `agent.ask` — Ask an agent | `agentId`, `input`, `outputType`, `output` |
+| `agent.run_command` — Run an agent command | `agentId`, `commandId`, `input` |
+| `agent.add_knowledge` — Add knowledge to an agent | `agentId`, `knowledgeType`, `title`, `knowledge` |
+| `agent.add_knowledge.v2` — Add knowledge to an agent (v2) | `agentId`, `knowledgeType`, `title`, `knowledge` |
+| `agent.add_project_as_knowledge` — Add a project to an agent | `agentId`, `projectId` |
+| `agentTeam.ask` — Ask an agent team | `agentTeamId`, `executionMode`, `input`, `output` |
### Schedule
@@ -68,7 +70,7 @@ Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub a
| Trigger | Properties |
| --- | --- |
-| `schedule.minutes` — Every X Minutes (Scheduled) — sub-hourly polling, paid-plan gated | `intervalMinutes` (5 / 10 / 15 / 30) |
+| `schedule.minutes` — Every X Minutes (Scheduled). Runs more than one time each hour. Needs a paid plan | `intervalMinutes` (5 / 10 / 15 / 30) |
| `schedule.hourly` — Every Hour (Scheduled) | - |
| `schedule.daily` — Every Day (Scheduled) | `timeOfDay`, `timezone` |
| `schedule.weekly` — Every Week (Scheduled) | `dayOfWeek`, `timeOfDay`, `timezone` |
@@ -80,12 +82,12 @@ Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub a
| Action | Properties |
| --- | --- |
-| `web.search` — Search Web | `query` |
-| `website.extract` — Scrape Webpage | `websiteUrl`, `errorBehavior` |
-| `pdf.convertToText` — Convert File to Text | `pdfUrl` |
-| `youtube.transcribeVideo` — Transcribe YouTube Video | `youtubeUrl` |
-| `url.addToMedia` — Upload File to Media | `url`, `emptyUrlBehavior` |
-| `text.addToMedia` — Upload Text to Media | `text`, `title` |
+| `web.search` — Search the web | `query` |
+| `website.extract` — Read a web page | `websiteUrl`, `errorBehavior` |
+| `pdf.convertToText` — Turn a file into text | `pdfUrl` |
+| `youtube.transcribeVideo` — Transcribe a YouTube video | `youtubeUrl` |
+| `url.addToMedia` — Save a file to Media | `url`, `emptyUrlBehavior` |
+| `text.addToMedia` — Save text to Media | `text`, `title` |
### Utilities
@@ -93,7 +95,7 @@ Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub a
| Action | Properties |
| --- | --- |
-| `transform.arrayIntoString` — Transform Array Into String | `transformEachItemIntoString`, `joinMechanism` |
+| `transform.arrayIntoString` — Transform Array to String | `transformEachItemIntoString`, `joinMechanism` |
## Communication & Collaboration
@@ -494,26 +496,26 @@ Premium integrations (for example Slack, Gmail, Google Sheets, HubSpot, GitHub a
* **Forms**: [AI Forms Documentation](https://www.taskade.com/learn/automation/forms-trigger)
* **Webhooks**: [Webhooks Documentation](https://www.taskade.com/learn/automation/webhook-trigger)
-* **Agent Triggers**: Built-in agent conversation triggers
+* **Agent Triggers**: built-in triggers that start from an agent conversation
### Control Actions
-* **Delay**: Wait/Sleep actions for timing control
-* **Branch**: [Conditional logic documentation](https://www.taskade.com/learn/automation/branch-action)
-* **Loop**: [Iteration control documentation](https://www.taskade.com/learn/automation/loop-action)
-* **Filter Data**: [Data filtering documentation](https://www.taskade.com/learn/automation/filter-data)
+* **Delay**: stops the automation for a time, then continues
+* **Branch**: sends the run down different paths ([Branch guide](https://www.taskade.com/learn/automation/branch-action))
+* **Loop**: runs steps one time for each item in a list ([Loop guide](https://www.taskade.com/learn/automation/loop-action))
+* **Filter Data**: lets the run continue only when the data meets a condition ([Filter Data guide](https://www.taskade.com/learn/automation/filter-data))
## Getting Started
-To learn more about setting up and using these integrations:
+For more about how to connect and use these integrations, read these pages:
-1. [**Getting Started with Automation**](https://www.taskade.com/learn/automation) - Complete beginner's guide
-2. [**AI Automation Collection**](https://www.taskade.com/learn/automation) - All automation resources
-3. [**Developer API**](../../apis-living-system-development/developers/api.md) - Build custom integrations
+1. [**Getting Started with Automation**](https://www.taskade.com/learn/automation): a guide for new users
+2. [**AI Automation Collection**](https://www.taskade.com/learn/automation): all automation guides
+3. [**Developer API**](../../apis-living-system-development/developers/api.md): build your own integrations
***
-_This reference is updated regularly with new integrations and features. For the latest updates, visit our_ [_Changelog_](../../changelog/)_._
+_We add new integrations and features to this reference. For the newest changes, read the_ [_Changelog_](../../changelog/)_._
---
diff --git a/genesis-living-system-builder/automation/integrations.md b/genesis-living-system-builder/automation/integrations.md
index c570bb1..1895c0a 100644
--- a/genesis-living-system-builder/automation/integrations.md
+++ b/genesis-living-system-builder/automation/integrations.md
@@ -5,24 +5,32 @@ description: >-
# Supported Integrations
-Taskade connects to **100+ tools and services** across communication, productivity, AI, and developer ecosystems. Integrations run inside automations: triggers (events that start automations) and actions (tasks that automations can perform). Some connectors offer only actions or only triggers.
+You can connect Taskade automations to **100+ tools and services** for communication, productivity, AI, and development. Each integration gives you triggers, actions, or both. A trigger is an event that starts an automation. An action is a step that the automation does. Some connectors have only actions or only triggers.
{% hint style="info" %}
-**Plans:** Premium integrations need a paid plan. Pro integrations need Pro and up. The Free plan runs the built-in Taskade steps and the standard connectors: HTTP Request, RSS, Pushover and MCP Client. The step picker shows the plan badge on each step.
+**Plans:** Premium integrations need a paid plan. Pro integrations need Pro and up. The Free plan runs the built-in Taskade steps and the standard connectors: HTTP Request, RSS, Pushover, and MCP Client. The step picker shows the plan badge on each step.
{% endhint %}
-## Enhanced Integration Features
+## What Integrations Give You
-### **Enterprise-Grade Reliability**
-- **Run history** for every automation, with the input and output of each step in the **Runs** tab
-- **Error handling** per step: **Continue if this step fails**, with **Stop here** or **Go to the next step**
-- **Secure connections**: you connect each account once with **Connect**, and every step in the workspace can reuse the connection
+### **Reliable Runs**
-### **Advanced Integration Capabilities**
-- **Data transformation** with utility actions that use no AI credits
-- **Complex multi-step workflows** spanning multiple services, with Branch, Filter Data, and Loop
-- **File processing** with Media actions and Convert File to Text
-- **Calendar integration** with Google Calendar, Microsoft Outlook Calendar, Cal.com, Calendly, and iCal feeds
+Each automation has these controls:
+
+- **Run history**: the **Runs** tab shows the input and the output of each step.
+- **Continue if this step fails**: you select **Stop here** or **Go to the next step** for each step.
+- **Retry if this step fails**: the step tries again when a connected app is busy or down for a few minutes.
+- **Wait for approval**: the run stops until a teammate approves or rejects it.
+- **Connections**: you connect each account one time with **Connect**. Every step in the workspace can use that connection again.
+
+### **What You Can Build**
+
+You can use integrations for these tasks:
+
+- **Change data**: utility actions change the format of data and use no AI credits.
+- **Connect many services in one automation**: use Branch, Filter Data, and Loop to control the flow.
+- **Read files**: use Media actions and Turn a file into text.
+- **Work with calendars**: connect Google Calendar, Microsoft Outlook Calendar, Cal.com, Calendly, and iCal feeds.
## Categories
@@ -110,31 +118,31 @@ Taskade connects to **100+ tools and services** across communication, productivi
| **Pushover** | — | Send a push notification | Every plan |
| **Forms** | Form submitted | — | Every plan |
-> **Related developer docs:** Build custom integrations programmatically with the [Developer Hub](../../apis-living-system-development/developer-home.md) and [Webhooks](../../apis-living-system-development/webhooks.md), or contribute to the open-source [Integration Kit](https://github.com/taskade/integrations) (Zapier & n8n actions and triggers).
+> **Related developer docs:** To build an integration in code, read the [Developer Hub](../../apis-living-system-development/developer-home.md) and [Webhooks](../../apis-living-system-development/webhooks.md). You can also add to the open-source [Integration Kit](https://github.com/taskade/integrations) (Zapier & n8n actions and triggers).
-## Integration Setup
+## Connect and Test an Integration
-### 1. **Authentication**
-Most integrations require authentication to access your accounts:
-- **OAuth 2.0**: Secure authorization flow (recommended)
-- **API Keys**: Simple token-based authentication
-- **Personal Access Tokens**: User-specific tokens with limited scope
+### 1. **Sign In to the Service**
+Most integrations need access to your account in the other service. The service uses one of these sign-in types:
+- **OAuth 2.0**: you sign in to the service in a browser window. Use this type when the service offers it.
+- **API key**: you paste a key from the service.
+- **Personal access token**: you paste a token that belongs to your user in the service.
-### 2. **Configuration**
-Each integration has specific configuration options:
-- **Endpoints**: Which API endpoints to use
-- **Data Mapping**: How to map fields between systems
-- **Rate Limits**: Respect service limitations
-- **Error Handling**: What to do when requests fail
+### 2. **Set the Step Fields**
+Each integration step has its own fields. Decide on these items:
+- **Endpoint**: the part of the service that the step calls.
+- **Field mapping**: which data goes into which field of the other service.
+- **Rate limits**: how many calls the service accepts in a period.
+- **Failures**: what the step does when a call fails.
-### 3. **Testing**
-Before deploying automations:
-- Test with sample data
-- Verify authentication works
-- Check data mapping accuracy
-- Monitor for errors and edge cases
+### 3. **Test the Automation**
+Do these steps before you turn the automation on:
+1. Run the automation with sample data.
+2. Make sure that the connection works.
+3. Make sure that each field gets the correct data.
+4. Read the **Runs** tab for errors.
-## Popular Integration Patterns
+## Common Integration Patterns
### **CRM Sync**
```
@@ -164,39 +172,39 @@ Google Sheets Update → AI Data Enrichment → Airtable Record Creation → Rep
## Advanced Features
### **Conditional Routing**
-Route data to different services based on content, priority, or other criteria.
+Send data to different services by its content, its priority, or other conditions. Use a **Branch** step for this.
### **Data Transformation**
-Clean, format, and enrich data as it flows between systems.
+Clean the data, change its format, and add data to it between two services.
### **Error Recovery**
-Choose what happens when a step fails: stop the run, skip the item, or go to the next step. Read the reason in the **Runs** tab.
+Select what happens when a step fails: stop the run, skip the item, or go to the next step. The **Runs** tab shows the reason for the failure.
### **Rate Limiting**
-Add a **Delay** step to space out calls to a service with rate limits.
+If a service limits the number of calls, add a **Delay** step between the calls.
### **Batch Processing**
-Process multiple records with **Loop**, or write several rows at once with Google Sheets **Add or Update Multiple Rows**.
+Use **Loop** to process many records. To write many rows in one step, use Google Sheets **Add or Update Multiple Rows**.
## Custom Integrations
-Don't see your service listed? You can create custom integrations using:
+If your service is not in the list, you can connect it with these tools.
### **HTTP Webhooks**
-- Send data to any REST API
-- Receive webhooks from external services (Pro and up)
-- Custom authentication headers
-- JSON, form data, raw, XML, and multipart file bodies
+- Send data to any REST API.
+- Receive webhooks from other services (Pro and up).
+- Add your own authentication headers.
+- Send a body as JSON, form data, raw text, XML, or multipart files.
### **Database Connections**
-- Use the **Supabase** connector (Pro and up) for a Postgres database
-- Call any other database through its HTTP API with **Send HTTP Request**
+- Connect a Postgres database with the **Supabase** connector (Pro and up).
+- Call another database through its HTTP API with **Send HTTP Request**.
### **File Processing**
-- Read uploaded files with **Convert File to Text**
-- Convert CSV to a list and a list to CSV with utility actions
-- Convert images and video with Media actions
-- Use Google Drive for cloud storage
+- Read uploaded files with **Turn a file into text**.
+- Change CSV into a list, and a list into CSV, with utility actions.
+- Convert images and video with Media actions.
+- Keep files in the cloud with Google Drive.
## Detailed Integration Examples
@@ -205,61 +213,61 @@ Don't see your service listed? You can create custom integrations using:
**Automatic Event Creation:**
- **Trigger:** Task with due date added
- **Action:** Create Google Calendar event
-- **Use Case:** Never miss project deadlines
+- **Use Case:** Show each project deadline in your calendar.
**Meeting Follow-up:**
- **Trigger:** New event in Google Calendar
- **Action:** Create follow-up tasks in Taskade
-- **Use Case:** Ensure meeting action items are tracked
+- **Use Case:** Keep a task for each action item from a meeting.
**Project Milestone Sync:**
- **Trigger:** Project completed
- **Action:** Create a calendar event for the next milestone
-- **Use Case:** Keep stakeholders informed of progress
+- **Use Case:** Tell stakeholders about the progress of the project.
### Google Sheets Integration
**Data Collection:**
- **Trigger:** Form submission
- **Action:** Add row to Google Sheets
-- **Use Case:** Centralized data tracking and analysis
+- **Use Case:** Keep all form data in one sheet for analysis.
**Inventory Management:**
- **Trigger:** Task completed (inventory check)
- **Action:** Update stock levels in spreadsheet
-- **Use Case:** Real-time inventory tracking
+- **Use Case:** Keep the stock levels current.
**Performance Reporting:**
- **Trigger:** Weekly schedule
-- **Action:** Generate report from Taskade data to Sheets
-- **Use Case:** Automated team performance dashboards
+- **Action:** Write a report from Taskade data to Sheets
+- **Use Case:** Keep a dashboard of team performance up to date.
### Social Media Integrations
**LinkedIn Content Sharing:**
- **Trigger:** Content approved in review project
- **Action:** Post to LinkedIn company page
-- **Use Case:** Automated content distribution
+- **Use Case:** Publish approved content with no manual step.
**Twitter/X Publishing:**
- **Trigger:** Task completed in the content calendar project
- **Action:** Create post on X/Twitter
-- **Use Case:** Publish approved posts on time
+- **Use Case:** Publish approved posts on time.
**YouTube Content Workflow:**
- **Trigger:** New video on your channel
- **Action:** Create promotion tasks and schedule posts
-- **Use Case:** Comprehensive video marketing automation
+- **Use Case:** Promote each new video.
### WhatsApp Business Integration
-WhatsApp Business integration enhances your communication strategy by automating customer interactions, managing inquiries, and enabling AI-powered responses.
+With WhatsApp Business, your automations can answer customers, log their questions, and send AI replies.
-#### **Setup & Configuration**
-1. **Connect WhatsApp Business Account** - Link your verified business account
-2. **Configure System User Access Token** - Set up API authentication
-3. **Verify Business Account ID** - Ensure proper account linking
-4. **Test Connection** - Send test message to verify setup
+#### **Connect WhatsApp Business**
+1. Connect your verified WhatsApp Business account.
+2. Paste your token in the **System User Access Token** field.
+3. Make sure that the **WhatsApp Business Account ID** is correct.
+4. Send a test message to make sure that the connection works.
#### **Available Actions**
@@ -274,32 +282,34 @@ WhatsApp Business integration enhances your communication strategy by automating
**Customer Support Automation:**
- **Trigger:** WhatsApp message received
-- **Action:** Create support ticket, auto-reply with case number
-- **AI Enhancement:** Sentiment analysis, automatic categorization
-- **Follow-up:** Route to appropriate specialist based on message content
+- **Action:** Create a support ticket and reply with the case number.
+- **AI step:** AI finds the mood of the message and puts it in a category.
+- **Next step:** Send the ticket to the correct specialist for the message content.
**Order Management:**
- **Trigger:** Stripe payment or webhook from your order system
-- **Action:** Send WhatsApp notification with tracking info
-- **Personalization:** Include customer name, order details, estimated delivery
-- **Escalation:** Alert if delivery is delayed
+- **Action:** Send a WhatsApp message with the tracking information.
+- **Personal details:** Add the customer name, the order details, and the expected delivery date.
+- **Escalation:** Send an alert if the delivery is late.
**Lead Qualification:**
- **Trigger:** WhatsApp message from new contact
-- **Action:** Run AI agent to qualify lead, update CRM
-- **Automation:** Schedule follow-up, assign to sales rep
-- **Integration:** Sync with HubSpot, Salesforce, or other CRM
+- **Action:** An AI agent qualifies the lead and updates the CRM.
+- **Next step:** Schedule a follow-up and assign it to a sales rep.
+- **CRM:** Sync the lead with HubSpot, Salesforce, or another CRM.
**Appointment Reminders:**
-- **Trigger:** Daily schedule that finds tomorrow's appointments
-- **Action:** Send WhatsApp reminder with appointment details
-- **Options:** Include reschedule link and location
-- **Confirmation:** Send a Request Approval message to confirm the time
+- **Trigger:** Daily schedule that finds the appointments for tomorrow
+- **Action:** Send a WhatsApp reminder with the appointment details.
+- **Options:** Add a link to change the time, and the location.
+- **Confirmation:** Send a Request Approval message to confirm the time.
---
## **Enterprise Integration Patterns**
+The examples below give the steps of each automation in outline.
+
### **Advanced CRM Workflows**
**HubSpot Revenue Operations**
@@ -524,17 +534,17 @@ Onboarding Automation:
## Getting Help
### **Documentation**
-Each integration has detailed setup guides and examples in our help center.
+Each integration has a setup guide and examples in the help center.
### **Support**
-Contact support for help with an integration. For Enterprise plans, contact sales.
+For help with an integration, contact support. For Enterprise plans, contact sales.
---
-> **Need a custom integration?**
-> → [Contact our team](https://taskade.com/contact) for enterprise solutions
+> **Do you need a custom integration?**
+> → [Contact our team](https://taskade.com/contact) about enterprise solutions
-> **Want to contribute an integration?**
+> **Do you want to add an integration?**
> → Open a PR on the open-source [Integration Kit](https://github.com/taskade/integrations) (Zapier & n8n actions and triggers)
→ **Next: [Browse Actions & Triggers](./actions.md)**
diff --git a/genesis-living-system-builder/automation/workflow-generator.md b/genesis-living-system-builder/automation/workflow-generator.md
index 74775bd..d29d132 100644
--- a/genesis-living-system-builder/automation/workflow-generator.md
+++ b/genesis-living-system-builder/automation/workflow-generator.md
@@ -1,6 +1,6 @@
---
description: >-
- Build complex automations in Taskade from plain English: the Workflow Generator creates triggers, branches, loops, and 100+ integrations with no drag-and-drop.
+ Build Taskade automations from plain English. The Workflow Generator creates triggers, branches, loops, and steps for 100+ integrations.
---
# Workflow Generator
@@ -9,21 +9,21 @@ description: >-
## Overview
-The Workflow Generator builds complex automation flows from **natural language**. Describe what you want in plain English, and the AI creates complete workflows with triggers, conditions, branches, loops, and integrations across 100+ services — no drag-and-drop, no manual configuration.
+With the Workflow Generator, you describe an automation in plain English. The AI then builds the full flow for you. The flow can have triggers, conditions, branches, loops, and steps for 100+ integrations. You do not drag steps or configure each step by hand.
-> **Taskade Genesis connection:** Taskade Genesis apps can include workflow automations as part of the Execution pillar. The Workflow Generator lets you create additional automations or modify existing ones. Each automation adds **+5 points** to your [Intelligence Score](https://www.taskade.com/learn/genesis), up to 30 points.
+> **Taskade Genesis connection:** A Taskade Genesis app can include automations as part of its Execution pillar. Use the Workflow Generator to add automations or to change the automations that you have. Each automation adds to your [Intelligence Score](https://www.taskade.com/learn/genesis).
---
## What You Can Build
-| Category | Example Prompt | What's Created |
+| Category | Example Prompt | What the Generator Makes |
|---|---|---|
-| **Multi-step workflows** | "When a form is submitted, create a task, notify the team, and send follow-up emails" | Form trigger → Task creation → Slack notification → Gmail follow-up |
-| **Content pipelines** | "Monitor X for new content, summarize it, and create a new project for each item" | RSS trigger → AI summary → Project creation loop |
-| **Scheduled reports** | "Every Monday, generate a progress report and share with stakeholders" | Weekly schedule → AI report generation → Email distribution |
+| **Multi-step workflows** | "When a form is submitted, create a task, notify the team, and send follow-up emails" | Form trigger → Add task → Slack message → Gmail follow-up |
+| **Content pipelines** | "Monitor X for new content, summarize it, and create a new project for each item" | RSS trigger → AI summary → Loop that creates one project for each item |
+| **Scheduled reports** | "Every Monday, generate a progress report and share with stakeholders" | Weekly schedule → AI report → Email to each stakeholder |
| **Agent-powered flows** | "Create an AI Agent that processes customer feedback and assigns follow-up" | Agent trigger → Sentiment analysis → Task assignment |
-| **Social distribution** | "Build an automation that posts our blog content to social media" | RSS/webhook trigger → AI rewrite → Multi-platform posting |
+| **Social distribution** | "Build an automation that posts our blog content to social media" | RSS or webhook trigger → AI rewrite → Posts on more than one platform |
---
@@ -31,40 +31,41 @@ The Workflow Generator builds complex automation flows from **natural language**
### Step-by-Step Tutorial
-1. Navigate to **Automations** tab → **Add automation** → **Generate With AI** (or visit `taskade.com/create/flow`)
-2. Describe your desired workflow and pick a mode (Auto, Agent Tool, Workflow, or Form)
-3. Watch the Workflow Generator build, refine, and optimize it
-4. Review the generated flow
-5. Click **Create Automation** to save it
-6. Customize triggers, actions, and conditions as needed
-7. Toggle the automation **On**
+1. Open the **Automations** tab. Click **Add automation**, then click **Generate With AI**. You can also go to `taskade.com/create/flow`.
+2. Describe the workflow that you want.
+3. Select a mode: Auto, Agent Tool, Workflow, or Form. Then press **Enter**.
+4. Wait until the Workflow Generator finishes the flow.
+5. Examine the flow that it made.
+6. Click **Create Automation** to save the flow.
+7. Change the triggers, actions, and conditions if necessary.
+8. Set the automation to **On**.
---
## Generator Modes
-| Mode | Description | Best For |
+| Mode | What It Does | Use It For |
|---|---|---|
-| **Auto** (default) | AI selects the optimal mode for your task | Most use cases — let AI decide |
-| **Agent Tool Mode** | Generates intelligent workflows as an AI agent tool | Agent-powered decision-making |
-| **Workflow Mode** | Sets up standard project automations and flows | Task management, notifications |
-| **Form Mode** | Generates smart forms that trigger actions on submission | Lead capture, surveys, intake |
+| **Auto** (default) | Selects the best mode for your prompt | Most tasks, when you are not sure which mode to use |
+| **Agent Tool Mode** | Makes an automation that an AI agent can use as a tool | Decisions that an agent makes |
+| **Workflow Mode** | Makes a standard automation | Task management and notifications |
+| **Form Mode** | Makes a form. A form submission starts the automation | Lead capture, surveys, and intake |
---
## Smart Edits & Chat Commands
-After generating a workflow, modify it with natural language:
+After the generator makes a workflow, you can change it with plain-English requests:
-| Edit Type | Example Prompt | What Happens |
+| Edit Type | Example Prompt | Result |
|---|---|---|
-| **Add step** | "Add a Slack notification to this workflow" | New Slack action added |
-| **Add logic** | "Add branching logic based on the form response" | Conditional branch inserted |
-| **Change trigger** | "Change the trigger to run daily instead of weekly" | Schedule trigger updated |
-| **Add loop** | "Loop through all tasks and send a reminder for each" | Loop action added |
-| **Modify action** | "Change the email to send to managers instead" | Action recipient updated |
-| **Add integration** | "Also post a summary to our Slack channel" | New integration step added |
-| **Optimize** | "Make this workflow more efficient" | AI suggests improvements |
+| **Add step** | "Add a Slack notification to this workflow" | The generator adds a Slack step |
+| **Add logic** | "Add branching logic based on the form response" | The generator adds a branch |
+| **Change trigger** | "Change the trigger to run daily instead of weekly" | The schedule trigger changes |
+| **Add loop** | "Loop through all tasks and send a reminder for each" | The generator adds a Loop step |
+| **Modify action** | "Change the email to send to managers instead" | The email goes to the new recipients |
+| **Add integration** | "Also post a summary to our Slack channel" | The generator adds a new integration step |
+| **Optimize** | "Make this workflow more efficient" | The AI gives suggestions to improve the flow |
---
@@ -72,35 +73,35 @@ After generating a workflow, modify it with natural language:
| Feature | What It Means |
|---|---|
-| **Instant Creation** | Build complex flows in minutes, not hours |
-| **Natural Language** | Describe any complexity — AI handles the implementation |
-| **Intelligent Editing** | Modify existing workflows with plain-English requests |
-| **Smart Recommendations** | AI-driven optimization tips and improvements |
-| **Built-in Integrations** | Connect with 100+ apps — you connect each account with **Connect**. Premium integrations need a paid plan |
-| **Branching** | Conditional paths based on data values |
-| **Looping** | Iterate over lists of items |
-| **Variables** | Dynamic data passing between steps (Liquid expressions) |
-| **AI Steps** | Agent-powered reasoning within workflows |
+| **Fast creation** | You get a complex flow in minutes |
+| **Natural Language** | You describe the flow in your words. The AI builds the steps |
+| **Plain-English edits** | You change a flow with a short request |
+| **Suggestions** | The AI suggests how to improve a flow |
+| **Built-in Integrations** | You can connect 100+ apps. Connect each account one time with **Connect**. Premium integrations need a paid plan |
+| **Branching** | A flow can follow different paths, based on the values in the data |
+| **Looping** | A Loop step runs its steps one time for each item in a list |
+| **Variables** | Steps send data to later steps with Liquid expressions |
+| **AI Steps** | An agent can make decisions inside the flow |
---
## Workflow Building Blocks
-| Block | Description | Example |
+| Block | What It Does | Example |
|---|---|---|
-| **Trigger** | What starts the flow | Form submitted, schedule, webhook, new email |
-| **Action** | What happens | Send email, create task, post to Slack |
-| **Branch** | Conditional logic (if/else) | If rating < 3, alert manager; else, log to database |
-| **Loop** | Iterate over a list | For each attachment, upload to Media |
-| **Delay** | Pause execution | Wait 24 hours, then send follow-up |
-| **AI Step** | Agent-powered processing | Agent categorizes ticket, generates response |
-| **Variable** | Dynamic data | `{{trigger.email}}`, `{{step1.output}}` |
+| **Trigger** | Starts the flow | A form submission, a schedule, a webhook, or a new email |
+| **Action** | Does the work | Send an email, create a task, or post to Slack |
+| **Branch** | Selects a path (if/else) | If the rating is less than 3, alert a manager. If not, log the result |
+| **Loop** | Runs its steps one time for each item | For each attachment, save the file to Media |
+| **Delay** | Stops the flow for a time | Wait 24 hours, then send a follow-up |
+| **AI Step** | Lets an agent do the work | An agent sorts a ticket and writes a reply |
+| **Variable** | Carries data between steps | `{{trigger.email}}`, `{{step1.output}}` |
---
## Example Prompts
-| Prompt | Flow Generated |
+| Prompt | Flow That the Generator Makes |
|---|---|
| "When a customer submits feedback with a rating below 3, alert the support manager on Slack and create a high-priority task" | Form Trigger → Branch (rating < 3) → Slack Message → Create Task (priority: high) |
| "Every day at 9 AM, summarize all overdue tasks and email them to team leads" | Daily Schedule → Find overdue tasks → AI summarize → Send Email |
@@ -110,12 +111,21 @@ After generating a workflow, modify it with natural language:
---
+## Use an Agent in a Generated Flow
+
+You can add a custom AI agent to a flow that the generator made:
+
+* Use the **Agent Tool** trigger. Then an agent can start the automation during a chat.
+* Add a **Run an agent command** step. Then the agent works at the exact step that you select.
+
+---
+
## What's Next
-| Guide | What You'll Learn |
+| Guide | What You Learn |
|---|---|
-| [Automation Triggers](./README.md) | All available triggers for your workflows |
-| [Taskade Autopilot](../ai-features/autopilot.md) | Generate entire workspace ecosystems |
+| [Automation Triggers](./README.md) | The triggers that you can use in your workflows |
+| [Taskade Autopilot](../ai-features/autopilot.md) | Generate a full workspace from a prompt |
---
diff --git a/genesis-living-system-builder/community-and-sharing/README.md b/genesis-living-system-builder/community-and-sharing/README.md
index 5a28a6b..a156974 100644
--- a/genesis-living-system-builder/community-and-sharing/README.md
+++ b/genesis-living-system-builder/community-and-sharing/README.md
@@ -5,132 +5,139 @@ description: >-
# Sharing: Living Systems
-**Turn your brilliant Taskade Genesis app into something that helps business owners everywhere. Share what you've built, discover amazing solutions from others, and copy complete apps with just one click.**
+**Share the Taskade Genesis apps that you build. Find apps from other builders, and copy a complete app with one click.**
{% hint style="success" %}
-**What you can do here.** Publish your Taskade Genesis apps to the community, let others fork them in one click with **Let others clone**, and sell original apps as paid App Kits.
+**What you can do here.** Publish your Taskade Genesis apps to the community. Let others copy them with **Let others clone**. Sell original apps as paid App Kits.
{% endhint %}
***
## **Publish and Share Your Apps**
-Share your Genesis creations with the world through publishing and cloning capabilities.
+You can publish your Taskade Genesis apps and let other people clone them.
### **Publishing Apps**
-Make your apps available to anyone with a link, creating public or private access points for your solutions.
+A published app has a link. You choose who can open it.
-**Publishing Benefits:**
+**What you get when you publish:**
-* **Direct Access**: Users can interact with your app through a dedicated URL
-* **No Installation Required**: Works immediately in any browser
-* **Version Control**: Update your published app while maintaining the same link
-* **Monetization Ready**: Sell original apps as paid App Kits with a one-time price
+* **Direct access**: Users open your app at its own URL.
+* **No installation**: The app works in any browser.
+* **Version control**: You can update your published app and keep the same link.
+* **Sales**: You can sell an original app as a paid App Kit for a one-time price.
-**Publishing Options:**
+**Publishing options:**
-* **Public Apps**: Accessible to anyone with the link. Discoverable in search only when **Show in public listings** is on
-* **Private Apps**: Require a shared password (Pro and up) or each person's own sign-in
+* **Public apps**: Anyone with the link can open them. Search shows them only when **Show in public listings** is on.
+* **Private apps**: Each person signs in, or types one shared password (Pro and up).
### **App Visibility Settings**
-Choose how people access your published app under **Who can open it** in the **Publish** panel:
+In the **Publish** panel, choose who can open your app under **Who can open it**.
-**Public Apps:**
+**Public apps:**
-* Anyone with the link can access and use your app
-* No authentication required
-* Ideal for marketing, demonstrations, or free tools
+* Anyone with the link can open and use your app.
+* Visitors do not sign in.
+* Good for marketing, demos, and free tools.
-**Private Apps:**
+**Private apps:**
-* Require each person to sign in (**Each person signs in**, on every plan), or one shared password (**Anyone with the password**, Pro and up)
-* Control who can use your app from the **App Users** tab
-* Suitable for client portals or internal tools
-* Sign-in confirms who a visitor is. It does not hide other people's records: turn on **Private entries** to show each person only their own rows
+* Each person signs in (**Each person signs in**, on every plan), or types one shared password (**Anyone with the password**, Pro and up).
+* You control who can use your app in the **App Users** tab.
+* Good for client portals and internal tools.
+* Sign-in confirms who a visitor is. It does not hide other people's records. To show each person only their own rows, turn on **Private entries**.
### **Unpublishing Apps**
-You can always unpublish your app if needed:
+You can unpublish your app at any time:
-1. **Open your Genesis app** in your workspace
-2. **Click Publish** in the top toolbar
-3. **Click the arrow on the Live button** at the bottom of the menu
-4. **Select "Unpublish app"**
+1. Open your Taskade Genesis app in your workspace.
+2. Click **Publish** in the top toolbar.
+3. Click the arrow on the **Live** button at the bottom of the menu.
+4. Click **Unpublish app**.
-**Note:** Unpublishing immediately removes public access, and the URL becomes inactive. Users who bookmarked the link will lose access.
+**Note:** Public access stops at once, and the URL stops working. Users who saved the link cannot open the app.
### **Enabling Cloning (Forking)**
-Allow others to create complete copies of your published app in their own workspaces. Open **Publish → Let others clone** and choose **Free** or **Paid** (**Off** stops new clones). A public listing alone does not make an app cloneable. (When others copy your public app it's called **forking**; **cloning** also refers to duplicating your own app inside your workspace with **Duplicate app**.)
+You can let other people copy your published app into their own workspaces:
-**Cloning Features:**
+1. Open **Publish → Let others clone**.
+2. Choose **Free** or **Paid**. **Off** stops new clones.
-* **One-Click Duplication**: Users can clone your entire app with a single click
-* **Complete Fidelity**: Projects, agents, automations, and media files are copied. Connected accounts and secrets are not
-* **Independent Operation**: Cloned apps function completely independently
-* **Customization Freedom**: Users can modify cloned apps without affecting the original
+A public listing alone does not make an app cloneable. When other people copy your public app, the copy is a **fork**. A **clone** can also mean a copy of your own app inside your workspace, made with **Duplicate app**.
+
+**What a clone gives:**
+
+* **One click**: Users clone your whole app with one click.
+* **Full copy**: Taskade copies the projects, agents, automations, and media files. Taskade does not copy connected accounts and secrets.
+* **Independent app**: The cloned app works on its own.
+* **Free changes**: Users can change their copy. The original app does not change.
### **When to Use Templates vs. Cloning**
-**Use Templates When:**
+**Use a template when:**
-* You want to provide starting points for customization
-* Users need to adapt the app significantly for their use case
-* You're sharing incomplete or example workflows
+* You give people a starting point that they change.
+* Users must change the app a lot for their use case.
+* You share an example workflow that is not finished.
-**Use Cloning When:**
+**Use a clone when:**
-* Your app is a complete, working solution
-* Users should have immediate functionality without setup
-* You want to share proven, production-ready workflows
+* Your app is a complete, working solution.
+* Users must get working functions with no setup.
+* You share a workflow that works in production.
## App Forking
-The most powerful way to learn and build is by starting with working solutions. Taskade Genesis lets you fork (copy) any community app whose creator turned on **Let others clone**.
+The fastest way to learn and build is to start with a working app. With Taskade Genesis, you can fork a community app when its creator turns on **Let others clone**.
### **What is App Forking?**
-Forking creates a complete, independent copy of a Genesis app in your workspace. Unlike simple templates, forking gives you:
+A fork is a complete, independent copy of a Taskade Genesis app in your workspace. A fork gives you more than a simple template:
-* **Complete Workspace Clone**: Projects, agents, automations, and media files
-* **Ready Workflows**: Automations arrive set up. Connect your own accounts, because connected accounts and secrets are not copied
-* **Full Customization Rights**: Modify, rebrand, and extend however you want
-* **Independent Evolution**: Your forked app develops separately from the original
+* **Full copy of the workspace parts**: projects, agents, automations, and media files
+* **Ready workflows**: the automations arrive set up. Connect your own accounts, because Taskade does not copy connected accounts and secrets.
+* **Full rights to change**: change, rebrand, and extend the app
+* **Independent copy**: your fork changes separately from the original
### **One-Click Complete Duplication**
-When you fork a Genesis app, you get:
+When you fork a Taskade Genesis app, you get:
-| What Gets Copied | Details |
-| ---------------------- | ----------------------------------------------------- |
-| **🎨 App Interface** | Complete layout, styling, and user experience |
-| **📊 Projects & Data** | Project structures and data (visitor submissions only if shipped as sample data) |
-| **🤖 AI Agents** | Trained agents with their knowledge and personalities |
-| **⚡ Automations** | Workflows (connect your own accounts) |
-| **📁 Media Files** | Images, documents, and assets |
-| **🔧 Configuration** | Settings, data access choices, and customizations |
+| What Gets Copied | Details |
+| --- | --- |
+| **🎨 App Interface** | The full layout, styles, and user experience |
+| **📊 Projects & Data** | Project structures and data. Visitor submissions stay out. A project shipped as sample data brings only the creator's own example rows |
+| **🤖 AI Agents** | Trained agents with their knowledge and personalities |
+| **⚡ Automations** | Workflows. Connect your own accounts |
+| **📁 Media Files** | Images, documents, and assets |
+| **🔧 Configuration** | Settings, data access choices, and customizations. **Invite only** sign-up and **Private entries** carry over |
+
+A project where each person sees only their own entries never ships its rows. Sign-in alone confirms who a visitor is. It does not hide other people's records. Test your copy with a second account before you share it.
### **Professional Use Cases**
-**For Agencies:**
+**For agencies:**
-* Fork client portal templates and customize for each client
-* Start with proven solutions instead of building from scratch
-* Maintain consistent quality across client projects
+* Fork a client portal template and change it for each client.
+* Start with a working solution, not a blank app.
+* Keep the same quality across client projects.
-**For Entrepreneurs:**
+**For entrepreneurs:**
-* Copy successful business models and adapt to your market
-* Learn from working examples of complex workflows
-* Accelerate time-to-market with proven foundations
+* Copy a working business model and change it for your market.
+* Learn from working examples of complex workflows.
+* Launch faster on a tested base.
-**For Teams:**
+**For teams:**
-* Share internal tools across departments
-* Standardize processes with working templates
-* Enable non-technical team members to deploy solutions
+* Share internal tools across departments.
+* Make processes the same with working templates.
+* Let team members who do not code put solutions to use.
***
@@ -138,73 +145,81 @@ When you fork a Genesis app, you get:
### What Happens When You Share?
-When you share your Genesis app with the community, it goes into a public library where other business owners can find it, try it out, and copy it for their own use. You can explore Taskade App Factory at [taskade.com/apps](https://www.taskade.com/apps) to see the shared apps other builders have published. Think of it like sharing a great recipe—you help others while showing off your skills!
+When you share your Taskade Genesis app with the community, it goes into a public library. Other business owners can find it, try it, and copy it. Explore the Taskade App Factory at [taskade.com/apps](https://www.taskade.com/apps) to see the apps that other builders published. Your shared app helps other people and shows your skills.
### When to Share with Community
-**Perfect for:**
+**Good to share:**
-* Universal business solutions (CRM templates, feedback systems, inventory trackers)
+* General business solutions (CRM templates, feedback systems, inventory trackers)
* Educational examples and tutorials
-* Industry-specific tools that benefit many organizations
-* Showcasing your development skills and creativity
-* Building your professional reputation
+* Industry tools that help many organizations
+* Apps that show your skills and ideas
+* Apps that build your professional reputation
-**Keep Private for:**
+**Keep private:**
-* Client-specific solutions with proprietary workflows
-* Apps containing sensitive business data or processes
-* Highly customized solutions that won't transfer well
-* Work-in-progress apps that aren't ready for public use
+* Client solutions with private workflows
+* Apps that contain sensitive business data or processes
+* Highly custom solutions that do not work for other people
+* Apps that are not ready for public use
***
## **Professional Publishing with Paywalls**
-Transform your Genesis apps into revenue-generating solutions with professional publishing features.
+You can sell your Taskade Genesis apps.
### **Monetize Your Expertise**
-Create premium Taskade Genesis apps that users pay for: open **Publish → Let others clone**, choose **Paid**, set a **one-time price**, and click **Put on sale**. Buyers pay to receive their own private copy of the app. Finish payout setup in **Earnings**. See [Paid Apps](publish-and-clone.md#paid-apps) for the full flow.
+To sell a Taskade Genesis app:
+
+1. Open **Publish → Let others clone**.
+2. Choose **Paid**.
+3. Set a **one-time price**.
+4. Click **Put on sale**.
+5. Finish the payout setup in **Earnings**.
+
+Buyers pay to get their own private copy of the app. Read [Paid Apps](publish-and-clone.md#paid-apps) for the full flow.
### **Perfect for:**
-**Consultants & Agencies:**
+**Consultants and agencies:**
* Client assessment tools and calculators
-* Industry-specific templates and workflows
-* Diagnostic and analysis applications
+* Industry templates and workflows
+* Diagnostic and analysis apps
-**SaaS Entrepreneurs:**
+**SaaS entrepreneurs:**
-* Niche business tools and utilities
-* Specialized dashboards and reporting
+* Small business tools for one market
+* Special dashboards and reports
* Process automation solutions
-**Content Creators:**
+**Content creators:**
* Educational tools and courses
* Interactive guides and tutorials
-* Community management platforms
+* Community management apps
### **Professional Features**
-| Feature | Benefit |
-| ------------------------- | ------------------------------------------------- |
-| **🔒 Access Control** | Buyers receive their own private copy of the app |
-| **📊 Analytics** | Built-in metrics on views, likes, and installs |
-| **💳 Payment Processing** | Checkout and payouts handled for you |
-| **🎨 Custom Branding** | Your logo, colors (Business+), and domain (Pro+) |
+| Feature | Benefit |
+| --- | --- |
+| **🔒 Access Control** | Buyers get their own private copy of the app |
+| **📊 Analytics** | Built-in counts of views, likes, and installs |
+| **💳 Payment Processing** | Taskade runs the checkout and the payouts |
+| **🎨 Custom Branding** | Your logo, colors (Business+), and domain (Pro+) |
***
## **Embeddable Public AI Agents**
-Share your AI expertise by embedding trained agents directly into websites, documentation, or any web page.
+You can put your trained agents into websites, documentation, or other web pages.
### **Agents Anywhere**
-The agent **Share** panel includes an embed snippet builder — copy a ready-made snippet (inline iframe or floating chat widget) and customize the **launcher icon** and background to match your site. The generated iframe points at your agent's public URL:
+The **Share** panel of an agent has an embed snippet builder. Copy a ready-made snippet: an inline iframe or a floating chat widget. You can change the **launcher icon** and the background to match your site. The iframe points at the public URL of your agent:
```html