> **Building with AI coding agents?** Install the authstack plugin with one command. This equips your agent with accurate Scalekit implementation patterns.
>
> **Recommended** (the `-y` flags skip prompts, so the command runs without an interactive terminal):
> ```bash
> npx -y @scalekit-inc/cli setup -y
> ```
>
> Global:
> ```bash
> npm install -g @scalekit-inc/cli
> scalekit setup -y
> ```
>
> Supports Claude Code, Cursor, GitHub Copilot, Codex + skills for other Agent Skills-compatible agents in `~/.agents/skills`.
> Skills: integrate-agentkit, implement-saaskit, add-mcp-oauth, implement-sso, implement-scim.
> [Full setup guide](https://docs.scalekit.com/dev-kit/build-with-ai/)

---

# Connect AI agents to CustomGPT AI

The CustomGPT AI connector lets your AI agent act in each user's CustomGPT AI account. Each user connects their own CustomGPT AI access token once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 59 tools.

**Authentication:** Bearer Token
**Categories:** AI
**Tools:** 59: 34 read, 19 write, 6 destructive
**Users sign in with:** Access token
**Built by:** Scalekit connector
**Try it:** [Playground in the Scalekit dashboard](https://app.scalekit.com/ws/signup?sk_intent=playground&provider=CUSTOMGPT)

## Setup

1. ### Install the SDK

   **Node.js**

   ```bash
   npm install @scalekit-sdk/node dotenv
   ```

   **Python**

   ```bash
   pip install scalekit-sdk-python python-dotenv
   ```

2. ### Set your credentials

   Add your Scalekit credentials to your `.env` file. Find values in **[app.scalekit.com](https://app.scalekit.com)** > **Developers** > **API Credentials**.

   ```sh title=".env"
   SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
   SCALEKIT_CLIENT_ID=<your-client-id>
   SCALEKIT_CLIENT_SECRET=<your-client-secret>
   ```

3. ### Create the CustomGPT AI connection

   In **AgentKit > Connections**, create a CustomGPT AI connection. The name you give it is the `connection_name` your code passes. See [Configure connections](/agentkit/connections/).

4. ### Authorize a user and make your first call

   **Node.js** (`quickstart.mts`)

   ```typescript
   import { ScalekitClient } from '@scalekit-sdk/node'
   import 'dotenv/config'
   import { createInterface } from 'node:readline/promises'

   const scalekit = new ScalekitClient(
     process.env.SCALEKIT_ENVIRONMENT_URL,
     process.env.SCALEKIT_CLIENT_ID,
     process.env.SCALEKIT_CLIENT_SECRET,
   )
   const actions = scalekit.actions

   const connector = 'customgpt'
   const identifier = 'user_123'

   // Generate an authorization link for the user
   const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
   console.log('Authorize CustomGPT AI:', link)
   const rl = createInterface({ input: process.stdin, output: process.stdout })
   await rl.question('Press Enter after authorizing...')
   rl.close()

   // Make your first call
   const result = await actions.executeTool({
     connector,
     identifier,
     toolName: 'customgpt_agents_list',
     toolInput: {},
   })
   console.log(result)
   ```

   ```bash
   npx tsx quickstart.mts
   ```

   **Python** (`quickstart.py`)

   ```python
   import os
   from scalekit import ScalekitClient
   from dotenv import load_dotenv
   load_dotenv()

   scalekit_client = ScalekitClient(
       env_url=os.getenv("SCALEKIT_ENVIRONMENT_URL"),
       client_id=os.getenv("SCALEKIT_CLIENT_ID"),
       client_secret=os.getenv("SCALEKIT_CLIENT_SECRET"),
   )
   actions = scalekit_client.actions

   connection_name = "customgpt"
   identifier = "user_123"

   # Generate an authorization link for the user
   link_response = actions.get_authorization_link(
       connection_name=connection_name,
       identifier=identifier,
   )
   print("Authorize CustomGPT AI:", link_response.link)
   input("Press Enter after authorizing...")

   # Make your first call
   result = actions.execute_tool(
       tool_input={},
       tool_name="customgpt_agents_list",
       connection_name=connection_name,
       identifier=identifier,
   )
   print(result)
   ```

   ```bash
   python quickstart.py
   ```

   Each user opens the link once and enters their CustomGPT AI credentials there. See [Authorize a user](/agentkit/tools/authorize/) for the full flow and statuses.

## Tools

Pass the exact name to `execute_tool`, with an input like each tool's example. To ask for a tool that's missing, use the [request form](https://scalekitsupport.portal.usepylon.com/forms/request-a-connector-tool).

**Node.js**

```typescript
const result = await actions.executeTool({
  toolName: 'customgpt_agent_get',
  toolInput: {
    project_id: 12345,
  },
  connector: 'customgpt',
  identifier: 'user_123',
})
```

**Python**

```python
result = actions.execute_tool(
    tool_name="customgpt_agent_get",
    tool_input={
        "project_id": 12345,
    },
    connection_name="customgpt",
    identifier="user_123",
)
```

`result.data` is the app's response as JSON, and `result.execution_id` (`executionId` in Node.js) is the ID of the call.

### `customgpt_agent_get`

Get Agent · Read-only

Get one agent's configuration and processing status.
Returns data (id, project_name, sitemap_path, is_chat_active, is_master, user_id, team_id, created_at, updated_at, deleted_at).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent to retrieve. Obtain agent IDs from the response of GET /api/v1/projects or POST /api/v1/projects. Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_agent_settings_get`

Get Agent Settings · Read-only

Get an agent's full settings: persona instructions, response behavior, citations, deployment options and feature flags.
Returns data (chatbot_avatar, chatbot_background_type, radius_styling, font_family, chatbot_background, chatbot_background_color, chatbot_color_scheme, default_prompt, example_questions, response_source).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_agent_stats_get`

Get Agent Stats · Read-only

Get an agent's usage statistics: conversations, queries, documents, words and crawl progress.
Returns data (pages_found, pages_crawled, pages_indexed, crawl_credits_used, query_credits_used, total_queries, total_words_indexed, total_storage_credits_used).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_agents_list`

List Agents · Read-only

List the agents (projects) in the account, newest first, optionally filtered by name.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use this to get the agent ID other tools need. Use customgpt_agent_get for one agent's status.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `name` (`string`, optional): Filter the results by agent name. Example: Support conversation
- `order` (`string`, optional): Sort direction for results. One of: asc, desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Field to sort results by. One of: id, created_at. Example: id. One of: `id`, `created_at`.
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1

Example input: `{}`

### `customgpt_child_agents_list`

List Child Agents · Read-only

List the child agents assigned to a Multi-Agent.
Returns data (children, total).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the Multi-Agent project. Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_citation_get`

Get Citation · Read-only

Get a citation's source title, description, URL and preview image.
Returns data (id, url, title, description, library_map_snapshot, image).
Get citation IDs from a message's citations.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `citation_id` (`integer`, required): The unique citation identifier. Example: 99001
- `project_id` (`integer`, required): The unique project identifier. Example: 12345

Example input: `{"citation_id":99001,"project_id":12345}`

### `customgpt_citation_preview_get`

Get Citation Preview · Read-only

Get a preview of the source document behind a citation.
Use after customgpt_citation_get.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique identifier of the citation source to preview. Obtained from citation objects returned by the messages endpoint. Example: 445566

Example input: `{"page_id":445566}`

### `customgpt_conversation_messages_list`

List Conversation Messages · Read-only

List the prompts and agent responses in a conversation, with citations and timestamps.
Returns data (conversation, messages).
Get the session ID from customgpt_conversations_list or customgpt_conversation_create.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `session_id` (`string`, required): The session ID of the converstaions to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77
- `include_insights` (`boolean`, optional): When true, includes customer_intelligence data in the response. Example: true
- `order` (`string`, optional): Sort direction for messages. One of: asc, desc. Example: desc. One of: `asc`, `desc`.
- `page` (`integer`, optional): Page number to return. Example: 1

Example input: `{"project_id":12345,"session_id":"<session_id>"}`

### `customgpt_conversations_list`

List Conversations · Read-only

List an agent's conversations, most recent first, optionally by user type, name or last activity.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use customgpt_conversation_messages_list for one conversation's messages. Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the conversations for. Example: 12345
- `last_updated_after` (`string`, optional): Only conversations that received message(s) after this date will be returned in results. If not set, the parameter is ignored. Example: 2026-01-01T00:00:00Z
- `name` (`string`, optional): Filter the results by conversation name. Example: Support conversation
- `order` (`string`, optional): Sort direction for results. One of: asc, desc. Example: desc. One of: `asc`, `desc`.
- `order_by` (`string`, optional): Field to sort results by. One of: id, created_at. Example: id. One of: `id`, `created_at`.
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1
- `user_filter` (`string`, optional): Filter conversations by user type. One of: all, anonymous, team_member, me. Example: all. One of: `all`, `anonymous`, `team_member`, `me`.

Example input: `{"project_id":12345}`

### `customgpt_deployment_settings_get`

Get Deployment Settings · Read-only

Get an agent's deployment settings for the embed, live chat, AI assistant, instant viewer and SGE widgets.
Returns data (embed, live_chat, ai_assistant, instant_viewer, sge).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_document_labels_get`

Get Document Tags · Read-only

Get the tags (labels) assigned to a document for access control.
Returns data (a list of records with id, name).
Get the page ID from customgpt_documents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique identifier of the page/document. Example: 445566
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_document_metadata_get`

Get Document Metadata · Read-only

Get a document's citation metadata: title, URL, description and image.
Returns data (id, url, title, description, library_map_snapshot, image).
Get the page ID from customgpt_documents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique page id. Example: 445566
- `project_id` (`integer`, required): The unique agent id. Example: 12345

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_documents_list`

List Documents · Read-only

List the documents (web pages and files) in an agent's knowledge base, with crawl and index status.
Returns data (project, pages).
Use this to find a page ID. Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique agent identifier. Example: 12345
- `crawl_status` (`string`, optional): Filter pages by crawl status. One of: all, ok, failed, n/a, queued, limited. Example: all. One of: `all`, `ok`, `failed`, `n/a`, `queued`, `limited`.
- `index_status` (`string`, optional): Filter pages by index status. One of: all, ok, failed, n/a, queued, limited. Example: all. One of: `all`, `ok`, `failed`, `n/a`, `queued`, `limited`.
- `limit` (`integer`, optional): Limit the number of documents to return. Defaults to 20. Example: 20
- `order` (`string`, optional): Sort direction for results. One of: asc, desc. Example: desc. One of: `asc`, `desc`.
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1
- `search` (`string`, optional): Filter documents by a search term. Performs a case-insensitive substring match against the document URL and filename. For example, search=annual-report matches both https://example.com/annual-report-2024.pdf and annual-report-summary.pdf. Supports wildcards: - * matches any sequence of characters (e.g., search=annual*report matches annual-sales-report.pdf and annual-lease-report.pdf) - ? matches a single character (e.g., search=v?-report matches v1-report, v2-report, v3-report) When omitted, all documents are returned. Example: annual-report

Example input: `{"project_id":12345}`

### `customgpt_licenses_list`

List Licenses · Read-only

List an agent's access licenses.
Returns data (a list of records with name, key, project_id, created_at, updated_at).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_message_claims_get`

Get Message Claims · Read-only

Get the factual claims extracted from a response and their verification against sources.
Returns data (extracted_claims, evaluation_duration_ms).
Run customgpt_message_verify first if no claims exist.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `prompt_id` (`integer`, required): The unique identifier of the message to retrieve the Verify Responses data for. Obtain message IDs from the messages list endpoint. Example: 678901
- `session_id` (`string`, required): The session ID of the conversation to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77

Example input: `{"project_id":12345,"prompt_id":678901,"session_id":"<session_id>"}`

### `customgpt_message_get`

Get Message · Read-only

Get one message with its prompt, response, citations and metadata.
Returns data (id, conversation_id, user_id, user_query, openai_response, citations, citations_external, citations_hidden, tools, metadata).
Get the prompt ID from customgpt_conversation_messages_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `prompt_id` (`integer`, required): The unique identifier of the message to retrieve. Obtain message IDs from GET /api/v1/projects/{projectId}/conversations/{sessionId}/messages. Example: 678901
- `session_id` (`string`, required): The session ID of the conversation to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77
- `include_insights` (`boolean`, optional): When true, includes customer_intelligence data in the response. Only has an effect if the lead capture feature is enabled for this agent. Example: true

Example input: `{"project_id":12345,"prompt_id":678901,"session_id":"<session_id>"}`

### `customgpt_message_trust_score_get`

Get Message Trust Score · Read-only

Get the trust score computed for a response.
Returns data (overall_status, stakeholder_analysis).
Run customgpt_message_verify first.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `prompt_id` (`integer`, required): The unique identifier of the message to retrieve the Verify Responses data for. Obtain message IDs from the messages list endpoint. Example: 678901
- `session_id` (`string`, required): The session ID of the conversation to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77

Example input: `{"project_id":12345,"prompt_id":678901,"session_id":"<session_id>"}`

### `customgpt_persona_versions_list`

List Persona Versions · Read-only

List the saved versions of an agent's persona instructions.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use customgpt_persona_version_activate to roll back. Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique agent identifier. Example: 12345
- `limit` (`integer`, optional): Limit the number of documents to return. Defaults to 20. Example: 20
- `page` (`integer`, optional): Page number to return. Example: 1

Example input: `{"project_id":12345}`

### `customgpt_report_analysis_get`

Get Analysis Report · Read-only

Get chart-ready time series of queries, conversations and queries per conversation, daily or weekly.
Returns data (queries, conversations, queries_per_conversation).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `filters` (`array`, required): Filters to return graph-ready analysis data for. Example: ["queries"]
- `project_id` (`integer`, required): The ID of the agent to get analysis for. Example: 12345
- `interval` (`string`, optional): Time interval for aggregating analytics data. One of: daily, weekly. Example: weekly. One of: `daily`, `weekly`.

Example input: `{"filters":["queries"],"project_id":12345}`

### `customgpt_report_conversations_get`

Get Conversations Report · Read-only

Get conversation metrics: total conversations and average queries per conversation.
Returns data (total, average_queries_per_conversation).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `filters` (`array`, required): Filters to return conversations metrics for. Example: ["total"]
- `project_id` (`integer`, required): The ID of the agent to get conversations for. Example: 12345

Example input: `{"filters":["total"],"project_id":12345}`

### `customgpt_report_intelligence_get`

Get Customer Intelligence · Read-only

Get AI-analyzed insights per query: intent, emotion, content source, verified claims score and risks, filterable by those fields.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the project to get customer intelligence for. Example: 12345
- `accuracy` (`string`, optional): Filter results by Verified Claims Score as range (format: min-max, e.g., 0-100) Example: 0-100
- `content_source` (`array`, optional): Filter results by Content Source (use repeated parameters, e.g. content_source[]=value) Example: ["in-context"]
- `end_date` (`string`, optional): End of the date range for the intelligence report. ISO 8601 datetime format (e.g., 2024-03-31T23:59:59Z). Data up to this point is included in the response. Example: 2026-01-31T23:59:59Z
- `limit` (`integer`, optional): Number of items to return per page. Maximum is 100. Example: 20
- `page` (`integer`, optional): Page number to return. Example: 1
- `request_source` (`array`, optional): Filter results by Deployment (use repeated parameters, e.g. request_source[]=value) Example: ["web"]
- `start_date` (`string`, optional): Start of the date range for the intelligence report. ISO 8601 datetime format (e.g., 2024-01-15T00:00:00Z). Data from this point onward is included in the response. Example: 2026-01-01T00:00:00Z
- `user_emotion` (`array`, optional): Filter results by User Emotion (use repeated parameters, e.g. user_emotion[]=value) Example: ["positive"]
- `user_intent` (`array`, optional): Filter results by User Intent (use repeated parameters, e.g. user_intent[]=value) Example: ["informational"]

Example input: `{"project_id":12345}`

### `customgpt_report_leads_get`

Get Leads Report · Read-only

Get leads (emails, names and custom fields) captured from conversations in a date range.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the project to get leads export for. Example: 12345
- `end_date` (`string`, optional): End date for the export range (ISO 8601 format) Example: 2026-01-31T23:59:59Z
- `limit` (`integer`, optional): Number of lead records to return per page. Maximum is 100. Example: 20
- `page` (`integer`, optional): Page number to return. Example: 1
- `session_id` (`string`, optional): Filter leads to a specific conversation by its session_id. When omitted, leads from all conversations in the project are returned. When the session_id does not match a conversation in this project the response is empty. Example: 12345
- `start_date` (`string`, optional): Start date for the export range (ISO 8601 format) Example: 2026-01-01T00:00:00Z

Example input: `{"project_id":12345}`

### `customgpt_report_queries_get`

Get Queries Report · Read-only

Get query metrics: total volume and query status breakdown.
Returns data (total, query_status).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `filters` (`array`, required): Filters to return queries metrics for. Example: ["total"]
- `project_id` (`integer`, required): The ID of the agent to get queries for. Example: 12345

Example input: `{"filters":["total"],"project_id":12345}`

### `customgpt_report_traffic_get`

Get Traffic Report · Read-only

Get conversation volume by deployment channel (web widget, API, Slack and others).
Returns data (sources).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `filters` (`array`, required): Filters to return traffic metrics for. Example: ["sources"]
- `project_id` (`integer`, required): The ID of the agent to get traffic for. Example: 12345

Example input: `{"filters":["sources"],"project_id":12345}`

### `customgpt_sources_list`

List Sources · Read-only

List an agent's data sources (sitemaps, uploads and integrations) with their sync settings.
Returns data (uploads, sitemaps, zapiers, confluences, youtubes, vimeos, podcasts, rsses, linkedins, sharepoints).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_tag_get`

Get Tag · Read-only

Get one tag with its name and assigned users.
Returns data (id, name, users).
Get the tag ID from customgpt_tags_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `label_id` (`integer`, required): The unique identifier of the tag to operate on. Example: 321
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345

Example input: `{"label_id":321,"project_id":12345}`

### `customgpt_tags_list`

List Tags · Read-only

List an agent's tags (labels) and the users assigned to each.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Tags control which users see answers from which documents. Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the project. Example: 12345
- `limit` (`integer`, optional): Maximum number of tags to return per page. Maximum is 100. Example: 20
- `page` (`integer`, optional): Page number to return. Example: 1

Example input: `{"project_id":12345}`

### `customgpt_task_get`

Get Task · Read-only

Poll a Plan & Act task for its status, new events and final answer.
Returns data (task, status, events, result).
Use after a task is started in CustomGPT.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `task_id` (`string`, required): The task ID. This is the value returned in data.id when you create a task via POST /tasks. Example: task_01HZX5K9ABC

Example input: `{"project_id":12345,"task_id":"<task_id>"}`

### `customgpt_tasks_list`

List Tasks · Read-only

List an agent's Plan & Act tasks, most recent first.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `limit` (`integer`, optional): Number of results per page. Example: 20
- `page` (`integer`, optional): The page to retrieve. Page numbering starts at 1. To retrieve all tasks, keep incrementing this value until it equals last_page in the response. Default: 1. Example: 1

Example input: `{"project_id":12345}`

### `customgpt_team_invitations_list`

List Team Invitations · Read-only

List pending invitations to the current team.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use customgpt_team_invitation_create to invite someone.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `limit` (`integer`, optional): Limit the number of results to return. Defaults to 20. Example: 20
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1

Example input: `{}`

### `customgpt_team_members_list`

List Team Members · Read-only

List the members of the current team with their roles.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use this to get user IDs for tag assignment.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `limit` (`integer`, optional): Limit the number of results to return. Defaults to 20. Example: 20
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1

Example input: `{}`

### `customgpt_teams_list`

List Teams · Read-only

List the teams the user belongs to.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Use this to see which team context the key works in.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `limit` (`integer`, optional): Limit the number of results to return. Defaults to 20. Example: 20
- `page` (`integer`, optional): Page number to retrieve. Page numbering starts at 1. Increment until the response's last_page value is reached to paginate through all results. Default: 1. Example: 1

Example input: `{}`

### `customgpt_usage_limits_get`

Get Usage Limits · Read-only

Get the account's current usage and plan limits for agents, words stored and monthly queries.
Returns data (max_projects_num, current_projects_num, max_total_storage_credits, current_total_storage_credits, max_queries, current_queries).
Use this before large uploads or query batches.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `customgpt_user_get`

Get Current User · Read-only

Get the profile of the user who owns the API key.
Returns data (created_at, email, id, current_team_id, name, profile_photo_url, updated_at).
Use this to confirm the connection.

Read only: yes. Destructive: no. Idempotent (safe to retry): yes.

Inputs: none.

Example input: `{}`

### `customgpt_agent_replicate`

Replicate Agent · Write

Create a full copy of an agent, including its knowledge base, persona and settings.
Returns data (id, project_name, sitemap_path, is_chat_active, is_master, user_id, team_id, created_at, updated_at, deleted_at).
Use this to make a test variant or template copy. Get the agent ID (projectId) from customgpt_agents_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `project_id` (`integer`, required): The unique agent identifier. Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_conversation_create`

Create Conversation · Write

Start a new, optionally named conversation session with an agent.
Returns data (created_at, updated_at, deleted_at, id, name, project_id, created_by, session_id).
Returns the session ID to use with customgpt_message_send. Get the agent ID (projectId) from customgpt_agents_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `name` (`string`, optional): Conversation name. Example: Support conversation

Example input: `{"project_id":12345}`

### `customgpt_conversation_update`

Update Conversation · Write

Rename a conversation.
Returns data (created_at, updated_at, deleted_at, id, name, project_id, created_by, session_id).
Get the session ID from customgpt_conversations_list or customgpt_conversation_create.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `session_id` (`string`, required): The unique session identifier. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77
- `name` (`string`, optional): The new name of the conversation. Example: Support conversation

Example input: `{"project_id":12345,"session_id":"<session_id>"}`

### `customgpt_deployment_settings_update`

Update Deployment Settings · Write

Update one deployment section (embed, live_chat, ai_assistant, instant_viewer or sge) of an agent. Only the fields sent change.
Returns data (embed, live_chat, ai_assistant, instant_viewer, sge).
Get the agent ID (projectId) from customgpt_agents_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique agent identifier. Example: 12345
- `section` (`string`, required): Deployment section to update. One of: embed, live_chat, ai_assistant, instant_viewer, sge. Example: live_chat. One of: `embed`, `live_chat`, `ai_assistant`, `instant_viewer`, `sge`.
- `auto_pop_up` (`integer`, optional): Live chat auto-popup target. Accepted values: 0 (disabled), 1 (desktop), 2 (mobile), 3 (desktop and mobile). Applies to: live_chat. One of: 0, 1, 2, 3. Example: 0. One of: `0`, `1`, `2`, `3`.
- `auto_prompt_starter_question` (`boolean`, optional): Automatically prompt the first starter question. Applies to: live_chat. Example: false
- `auto_start_conversation` (`boolean`, optional): Automatically start a new conversation when the widget opens. Applies to: live_chat. Example: true
- `chatbot_type` (`string`, optional): Visual style of the deployment widget. live_chat accepts window, compact, floating; ai_assistant and instant_viewer accept window, compact. Applies to: live_chat, ai_assistant, instant_viewer. One of: window, compact, floating. Example: compact. One of: `window`, `compact`, `floating`.
- `continue_conversation` (`boolean`, optional): Persist and continue the previous conversation. Applies to: live_chat. Example: false
- `custom_search_parameter` (`string`, optional): URL query parameter name to pass as the search query. Alphanumeric, dash, underscore; max 15 characters. Applies to: sge. Example: q
- `default_button_color` (`string`, optional): Hex color for the default trigger button. Accepted formats: #FFF, #FFFFFF, or #FFFFFFFF. Applies to: ai_assistant, instant_viewer. Example: #2563EB
- `default_button_text` (`string`, optional): Label text on the default trigger button. Applies to: ai_assistant, instant_viewer. Example: Ask AI
- `delay_iframe_load_ms` (`number`, optional): Milliseconds to delay iframe load. Applies to: embed, sge. Example: 0
- `do_not_prompt_on_mobile` (`boolean`, optional): Suppress auto-prompts on mobile devices. Applies to: live_chat. Example: false
- `height` (`number`, optional): Widget height value. Applies to: embed, live_chat, sge. Example: 500
- `height_type` (`string`, optional): Whether height is fixed or automatic. Applies to: embed, live_chat, sge. One of: fixed, auto. Example: fixed. One of: `fixed`, `auto`.
- `height_unit` (`string`, optional): Unit for height. Applies to: embed, live_chat, sge. One of: px, %. Example: px. One of: `px`, `%`.
- `hide_prompt_from_starter_question_list` (`boolean`, optional): Hide the pre-prompt from the starter question list. Applies to: live_chat. Example: false
- `icon_size` (`string`, optional): Icon size for the live chat bubble trigger. Applies to: live_chat. One of: small, medium, large. Example: small. One of: `small`, `medium`, `large`.
- `is_crm_integration_enabled` (`boolean`, optional): Whether CRM integration is enabled for this deployment. Applies to: embed, live_chat, ai_assistant, instant_viewer, sge. Example: false
- `is_custom_context_enabled` (`boolean`, optional): Whether custom context injection is enabled. Applies to: embed, live_chat, ai_assistant, instant_viewer, sge. Example: false
- `is_labels_filter_enabled` (`boolean`, optional): Filter responses by source labels. Requires a plan with source labeling. Applies to: embed, live_chat, ai_assistant, instant_viewer, sge. Example: false
- `is_pre_prompt_enabled` (`boolean`, optional): Whether the pre-prompt is enabled for this deployment. Applies to: embed, live_chat, ai_assistant, instant_viewer. Example: false
- `is_using_custom_element` (`boolean`, optional): Whether a custom HTML element is used as the trigger. Applies to: ai_assistant, instant_viewer. Example: false
- `is_using_default_button` (`boolean`, optional): Whether to show the default styled button. Applies to: ai_assistant, instant_viewer. Example: true
- `labels_exclusive` (`boolean`, optional): When label filtering is active, use exclusive AND matching. Applies to: embed, live_chat, ai_assistant, instant_viewer, sge. Example: false
- `mobile_font_size` (`boolean`, optional): Use mobile-optimized font sizing. Applies to: embed, live_chat, ai_assistant, instant_viewer, sge. Example: true
- `position` (`string`, optional): Screen position of the live chat bubble. Applies to: live_chat. One of: right, left. Example: right. One of: `right`, `left`.
- `reset_conversation` (`boolean`, optional): Reset the conversation on each new session. Applies to: embed, live_chat, ai_assistant, instant_viewer. Example: false
- `width` (`number`, optional): Widget width value. Applies to: embed, live_chat. Example: 100
- `width_unit` (`string`, optional): Unit for width. Applies to: embed, live_chat. One of: px, %. Example: %. One of: `px`, `%`.

Example input: `{"project_id":12345,"section":"embed"}`

### `customgpt_document_labels_set`

Set Document Tags · Write

Replace the tags (labels) on a document. An empty list makes it visible to everyone.
Returns data (a list of records with id, name).
Get tag IDs from customgpt_tags_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique identifier of the page/document. Example: 445566
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `label_ids` (`array`, optional): List of label IDs to set for this page (replaces existing labels) Example: [321]

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_document_metadata_update`

Update Document Metadata · Write

Update the title, URL, description or image shown when a document is cited.
Returns data (id, url, title, description, library_map_snapshot, image).
Get the page ID from customgpt_documents_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique page id. Example: 445566
- `project_id` (`integer`, required): The unique agent id. Example: 12345
- `description` (`string`, optional): The description of the document used in citations. Example: Current pricing and plan limits
- `image` (`string`, optional): The url of the image used in citations. Example: https://example.com/og.png
- `title` (`string`, optional): The title of the document used in citations. Example: Pricing FAQ
- `url` (`string`, optional): The url of the document used in citations. Example: https://example.com/pricing

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_document_reindex`

Reindex Document · Write

Re-crawl and re-index one web page document so the agent uses its current content.
Returns data (updated).
Only works for web pages, not uploaded files.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique identifier of the page to reindex. Obtain page IDs from GET /api/v1/projects/{projectId}/pages. Example: 445566
- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_message_feedback_update`

Update Message Feedback · Write

Record thumbs up, down or neutral feedback on a response.
Returns data (id, conversation_id, user_id, user_query, openai_response, citations, citations_external, citations_hidden, tools, metadata).
Get the prompt ID from customgpt_conversation_messages_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `prompt_id` (`integer`, required): The unique identifier of the message to retrieve. Obtain message IDs from GET /api/v1/projects/{projectId}/conversations/{sessionId}/messages. Example: 678901
- `session_id` (`string`, required): The session ID of the conversation to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77
- `include_insights` (`boolean`, optional): When true, includes customer_intelligence data in the response. Only has an effect if the lead capture feature is enabled for this agent. Example: true
- `reaction` (`string`, optional): The new reaction. One of: neutral, disliked, liked. Example: liked. One of: `neutral`, `disliked`, `liked`.

Example input: `{"project_id":12345,"prompt_id":678901,"session_id":"<session_id>"}`

### `customgpt_message_send`

Send Message · Write

Ask an agent a question using the OpenAI-compatible chat format and get its answer grounded in the agent's knowledge base. Each call uses one query from the plan.
Returns id, object, created, model, usage, choices.
Pass the conversation so far in messages, ending with the user's question. Streaming is turned off. Get the agent ID (projectId) from customgpt_agents_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `messages` (`array`, required): The messages to send to the agent. Example: [{"role": "user", "content": "What are your support hours?"}]
- `project_id` (`integer`, required): The unique identifier of the agent to send the message to. Example: 12345
- `external_id` (`string`, optional): External identifier. Example: crm-1001
- `is_inline_citation` (`boolean`, optional): Whether to include inline citations. Example: false
- `labels` (`array`, optional): Optional source-label filter. Accepts two formats: (1) Flat array of label IDs or names for OR semantics, e.g. ["A","B","C"]. (2) Array of arrays for AND/OR logic (inner=OR, outer=AND), e.g. [["A","B"],["C","D"]] means (A OR B) AND (C OR D). Max 20 groups, 50 labels per group, 500 total. Invalid values are ignored. Example: ["billing"]
- `labels_exclusive` (`boolean`, optional): If true, search only pages with provided labels. If false, search pages with provided labels plus unlabeled pages. Example: false
- `lang` (`string`, optional): Response language. Example: en
- `model` (`string`, optional): ID of the model to use. Example: example

Example input: `{"messages":[{"role":"user","content":"What are your support hours?"}],"project_id":12345}`

### `customgpt_message_verify`

Verify Message · Write

Start fact-checking a response's claims against the agent's sources. Uses verification credits.
Returns data (message).
Read results with customgpt_message_claims_get and customgpt_message_trust_score_get.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `project_id` (`integer`, required): The ID of the agent to get the messages for. Example: 12345
- `prompt_id` (`integer`, required): The unique identifier of the message to retrieve the Verify Responses data for. Obtain message IDs from the messages list endpoint. Example: 678901
- `session_id` (`string`, required): The session ID of the conversation to get the messages for. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77

Example input: `{"project_id":12345,"prompt_id":678901,"session_id":"<session_id>"}`

### `customgpt_persona_version_activate`

Activate Persona Version · Write

Make an earlier persona version the active persona (saved as a new version).
Returns data (updated).
Get the version from customgpt_persona_versions_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `version` (`integer`, required): The version of the persona to set active. Example: 10

Example input: `{"project_id":12345,"version":10}`

### `customgpt_source_sync`

Sync Source Now · Write

Re-sync a sitemap source immediately instead of waiting for its schedule.
Returns data (id, created_at, updated_at, type, sub_type, settings, pages).
Get the source ID from customgpt_sources_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent that owns this source. Example: 12345
- `source_id` (`integer`, required): The unique identifier of the source to sync immediately. Obtain source IDs from GET /api/v1/projects/{projectId}/sources. Example: 7788

Example input: `{"project_id":12345,"source_id":7788}`

### `customgpt_source_update`

Update Source · Write

Update a data source's refresh frequency, page handling, JavaScript rendering and image extraction.
Returns data (id, created_at, updated_at, type, sub_type, settings, pages).
Get the source ID from customgpt_sources_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `source_id` (`integer`, required): The unique identifier of the source to update. Example: 7788
- `allow_extra_timeout` (`boolean`, optional): Allow extra timeout for agent source. Example: false
- `create_new_pages` (`boolean`, optional): Add new pages to agent automatically during refresh agent source. Example: true
- `data_refresh_frequency` (`string`, optional): How often the source content is re-fetched and re-indexed. Values: - never: no automatic refresh. - daily / weekly / monthly: refresh on that cadence. - advanced: uses either custom schedule field. One of: never, daily, weekly, monthly, advanced. Example: never. One of: `never`, `daily`, `weekly`, `monthly`, `advanced`.
- `executive_js` (`boolean`, optional): Whether the agent source should execute JavaScript. Example: true
- `image_extraction_type` (`string`, optional): The method used for extracting images from the website. One of: none, sync_from_sitemap, extract_opengraph_images. Example: sync_from_sitemap. One of: `none`, `sync_from_sitemap`, `extract_opengraph_images`.
- `refresh_existing_pages` (`string`, optional): Refresh existing page during refresh agent source. One of: never, always, if_updated. Example: never. One of: `never`, `always`, `if_updated`.
- `remove_unexist_pages` (`boolean`, optional): Remove pages from agent automatically during refresh agent source. Example: false
- `sync_all_images` (`boolean`, optional): When extracting images from a sitemap, sync every image on a page instead of only the first one. Each synced image consumes one vision credit. Example: false
- `update_existing_images` (`boolean`, optional): Whether to update existing images during refresh. Example: false

Example input: `{"project_id":12345,"source_id":7788}`

### `customgpt_tag_update`

Rename Tag · Write

Rename a tag. Its user and document assignments are kept.
Returns data (id, name).
Get the tag ID from customgpt_tags_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `label_id` (`integer`, required): The unique identifier of the tag to operate on. Example: 321
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `name` (`string`, optional): New tag name (must not start with a digit) Example: Support conversation

Example input: `{"label_id":321,"project_id":12345}`

### `customgpt_tag_users_assign`

Assign Users to Tags · Write

Set the users assigned to each tag. The list sent replaces the current assignment.
Returns data (current_page, data, first_page_url, from, last_page, last_page_url, next_page_url, path, per_page, prev_page_url).
Users are given by email; get them from customgpt_team_members_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. You can find this in the CustomGPT dashboard or from the response of POST /api/v1/projects. Example: 12345
- `labels` (`array`, optional): Tags with the users to assign: each item has the tag id and a users list of user emails, for example [{"id": 321, "users": ["manager@example.com"]}]. Example: [{"id": 321, "users": ["manager@example.com"]}]

Example input: `{"project_id":12345}`

### `customgpt_tags_create`

Create Tags · Write

Create one or more tags for document access control.
Returns data (a list of records with id, name).
Assign them with customgpt_document_labels_set and customgpt_tag_users_assign.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. You can find this in the CustomGPT dashboard or from the response of POST /api/v1/projects. Example: 12345
- `names` (`array`, optional): Tag names to create/restore (max 20 characters each, must not start with a digit) Example: ["billing", "internal"]

Example input: `{"project_id":12345}`

### `customgpt_task_stop`

Stop Task · Write

Stop the response a running task is generating. Work already done is kept and billed.
Get the task ID from customgpt_tasks_list.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `task_id` (`string`, required): The task ID. Example: task_01HZX5K9ABC

Example input: `{"project_id":12345,"task_id":"<task_id>"}`

### `customgpt_task_update`

Update Task · Write

Rename a Plan & Act task.
Returns data (id, project_id, name, created_at, updated_at).
Get the task ID from customgpt_tasks_list.

Read only: no. Destructive: no. Idempotent (safe to retry): yes.

Inputs:

- `name` (`string`, required): New task name. Example: Support conversation
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `task_id` (`string`, required): The task ID. This is the value returned in data.id when you create a task via POST /tasks. Example: task_01HZX5K9ABC

Example input: `{"name":"<name>","project_id":12345,"task_id":"<task_id>"}`

### `customgpt_team_invitation_create`

Invite Team Member · Write

Email an invitation to join the current team with a role.
Returns data (id, email, role_id, team_id, created_at).
Get role IDs from the team's roles in CustomGPT. The recipient must accept.

Read only: no. Destructive: no. Idempotent (safe to retry): no.

Inputs:

- `email` (`string`, optional): Email address of the person to invite. Example: teammate@example.com
- `role_id` (`integer`, optional): Role ID to assign to the invited member. Example: 3

Example input: `{}`

### `customgpt_agent_delete`

Delete Agent · Destructive

Permanently delete an agent with its knowledge base, conversations, settings and analytics.
Returns data (deleted).
This cannot be undone. Get the agent ID (projectId) from customgpt_agents_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"project_id":12345}`

### `customgpt_conversation_delete`

Delete Conversation · Destructive

Permanently delete a conversation and all its messages.
Returns data (deleted).
Get the session ID from customgpt_conversations_list or customgpt_conversation_create.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `session_id` (`string`, required): The unique session identifier. Example: f1b9aa61-1c3a-4b3e-9d3c-2f6a0e1c9d77

Example input: `{"project_id":12345,"session_id":"<session_id>"}`

### `customgpt_document_delete`

Delete Document · Destructive

Remove a document from an agent's knowledge base.
Returns data (deleted).
Get the page ID from customgpt_documents_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `page_id` (`integer`, required): The unique identifier of the page (document) to delete. Obtain page IDs from GET /api/v1/projects/{projectId}/pages. Example: 445566
- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345

Example input: `{"page_id":445566,"project_id":12345}`

### `customgpt_source_delete`

Delete Source · Destructive

Delete a data source and the documents it added.
Returns data (deleted).
Get the source ID from customgpt_sources_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent (formerly known as project). Example: 12345
- `source_id` (`integer`, required): The unique identifier of a source. Example: 7788

Example input: `{"project_id":12345,"source_id":7788}`

### `customgpt_tag_delete`

Delete Tag · Destructive

Delete a tag and all its user and document assignments.
Returns data (deleted).
Documents left without tags become visible to everyone.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `label_id` (`integer`, required): The unique identifier of the tag to operate on. Example: 321
- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345

Example input: `{"label_id":321,"project_id":12345}`

### `customgpt_task_delete`

Delete Task · Destructive

Permanently delete a task and its messages.
Returns data (deleted).
Get the task ID from customgpt_tasks_list.

Read only: no. Destructive: yes. Idempotent (safe to retry): yes.

Inputs:

- `project_id` (`integer`, required): The unique identifier of the agent. Example: 12345
- `task_id` (`string`, required): The task ID. This is the value returned in data.id when you create a task via POST /tasks. Example: task_01HZX5K9ABC

Example input: `{"project_id":12345,"task_id":"<task_id>"}`

## Related

Other AI connectors ([all 216](/agentkit/connectors/?category=ai)):

- [Chorus](/agentkit/connectors/chorus/): Scalekit connector, Username and password, 9 tools
- [Gong](/agentkit/connectors/gong/): Scalekit connector, OAuth, 69 tools
- [Jiminny](/agentkit/connectors/jiminny/): Scalekit connector, Access token, 27 tools


---

## More Scalekit documentation

| Resource | What it contains | When to use it |
|----------|-----------------|----------------|
| [/llms.txt](/llms.txt) | Structured index with routing hints per product area | Start here — find which documentation set covers your topic before loading full content |
| [/llms-full.txt](/llms-full.txt) | Complete documentation for all Scalekit products in one file | Use when you need exhaustive context across multiple products or when the topic spans several areas |
| [sitemap-0.xml](https://docs.scalekit.com/sitemap-0.xml) | Full URL list of every documentation page | Use to discover specific page URLs you can fetch for targeted, page-level answers |
