> ## Documentation Index
> Fetch the complete documentation index at: https://docs.elementum.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Connect AI Agents to the Docs

> Give AI tools like Cursor, Claude, and ChatGPT real-time access to the Elementum documentation through our Model Context Protocol (MCP) server.

## Overview

The Elementum documentation is published with a built-in **Model Context Protocol (MCP) server**. MCP is an open standard that lets AI assistants — including Cursor, Claude, and ChatGPT — read a live source of information and use it to answer questions.

When you connect an AI tool to the Elementum docs MCP server, the assistant can:

* Search the full documentation in real time instead of relying on stale training data
* Return answers grounded in current Elementum guidance, with links back to the source pages
* Reference the docs while it helps you build apps, write automations, or troubleshoot issues

The server is hosted at:

```
https://docs.elementum.io/mcp
```

No account or credentials are required — the Elementum documentation is public, so any MCP-compatible client can connect.

***

## Connect an AI Tool

Pick the tool you use below. Each one asks for the same thing: a **name** (whatever you want to call it, for example `elementum-docs`) and the **URL** `https://docs.elementum.io/mcp`.

<Tabs>
  <Tab title="Cursor">
    1. Open the command palette with `Cmd/Ctrl + Shift + P` and search for **Open MCP Settings**.
    2. Select **Add custom MCP** to open `mcp.json`.
    3. Add the Elementum docs server to the configuration:

    ```json theme={null}
    {
      "mcpServers": {
        "elementum-docs": {
          "url": "https://docs.elementum.io/mcp"
        }
      }
    }
    ```

    4. Save the file. Cursor connects automatically and the new server appears in **MCP Settings**.
  </Tab>

  <Tab title="Claude Desktop">
    1. Open **Settings** in Claude Desktop and go to the **Connectors** page.
    2. Select **Add custom connector**.
    3. Enter the following:
       * **Name**: `Elementum Docs`
       * **URL**: `https://docs.elementum.io/mcp`
    4. Select **Add**. Claude connects to the server and the docs become available in new chats.
  </Tab>

  <Tab title="Claude Code">
    Run the following command in your terminal:

    ```bash theme={null}
    claude mcp add --transport http elementum-docs https://docs.elementum.io/mcp
    ```

    Claude Code registers the server and makes it available in your next session.
  </Tab>

  <Tab title="ChatGPT">
    1. In ChatGPT, open **Settings** → **Connectors** (available on paid plans that support custom MCP connectors).
    2. Select **Add custom connector** and enter:
       * **Name**: `Elementum Docs`
       * **URL**: `https://docs.elementum.io/mcp`
    3. Save the connector and enable it for the conversations where you want the assistant to consult the docs.
  </Tab>

  <Tab title="Other MCP Clients">
    Any client that supports the MCP **Streamable HTTP** transport can connect. Provide the server URL in whatever configuration format the client uses:

    ```json theme={null}
    {
      "mcpServers": {
        "elementum-docs": {
          "url": "https://docs.elementum.io/mcp"
        }
      }
    }
    ```

    Refer to your tool's documentation for the exact location of its MCP configuration file.
  </Tab>
</Tabs>

***

## Using the Docs From Your AI Tool

Once connected, ask the assistant questions as you normally would. When a question relates to Elementum, the assistant will query the MCP server and cite the relevant docs pages in its response.

Example prompts:

* "Using the Elementum docs, explain how assignment rules differ from approval processes."
* "Walk me through setting up a CloudLink to Snowflake. Link to the steps in the Elementum docs."
* "How do I build an automation that runs on a record update? Check the Elementum docs."

<Tip>
  If an answer looks out of date, ask the assistant to re-query the docs. MCP reads the live site, so updates to the documentation are reflected immediately.
</Tip>

***

## Troubleshooting

<AccordionGroup>
  <Accordion title="The AI tool can't reach the server">
    Confirm the URL is entered exactly as `https://docs.elementum.io/mcp` with no trailing slash or extra path segments. Check that your network allows outbound HTTPS requests to `docs.elementum.io`.
  </Accordion>

  <Accordion title="The assistant isn't using the docs in its answers">
    Most clients require you to explicitly mention the connector or include a phrase like "check the Elementum docs" in your prompt. Verify the connector is enabled for the current conversation in your tool's settings.
  </Accordion>

  <Accordion title="I want to remove the connection">
    Open the MCP settings in your tool and delete the `elementum-docs` entry (or whatever name you used). The assistant will stop querying the docs on its next turn.
  </Accordion>
</AccordionGroup>

***

## Related Resources

* [API Reference](/api-reference/api-introduction) — connect agents directly to Elementum's REST API
* [Agents Overview](/ai-agents/ai-overview) — build AI agents inside the Elementum platform
* [Help & Resources](/support/resources) — other support channels