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

# Use with AI (MCP)

> Connect the eBRC API docs to Claude, Cursor, or ChatGPT over MCP, call the API from your AI assistant with your own key, or feed any page to an LLM as clean Markdown.

These docs are built to be driven by AI. Add one MCP server and your assistant can search and answer from the whole knowledge base; open the API Reference and it can help you call the API with your own key; or hand any page to an LLM as clean Markdown.

## Connect over MCP

The docs are served as an MCP (Model Context Protocol) server. Point any MCP-capable tool at one URL and it can search and read this documentation from inside the tool.

```text theme={"dark"}
https://ebrcapi.docs.eximfiles.io/mcp
```

<Tabs>
  <Tab title="Claude Code">
    ```bash theme={"dark"}
    claude mcp add --transport http ebrc-docs https://ebrcapi.docs.eximfiles.io/mcp
    ```

    Then ask questions like "how do I submit IRMs for eBRC generation" and it answers from these docs, with citations.
  </Tab>

  <Tab title="Cursor">
    Add this to `.cursor/mcp.json` in your project (or `~/.cursor/mcp.json` globally):

    ```json theme={"dark"}
    {
      "mcpServers": {
        "ebrc-docs": {
          "url": "https://ebrcapi.docs.eximfiles.io/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="ChatGPT">
    In ChatGPT, open **Settings**, then **Connectors**, and add a custom connector with the URL `https://ebrcapi.docs.eximfiles.io/mcp`.
  </Tab>
</Tabs>

<Note>
  The docs MCP is **read-only**: it searches and returns documentation. It never sees, asks for, or transmits your API key. Calling the API is a separate step, below, and always uses your own key.
</Note>

## Ask AI on any page

Every page has a contextual menu next to its title. Use it to **copy the page as Markdown**, **view it as plain text**, or **open it in ChatGPT or Claude** with the page content already attached, so you can ask questions without leaving your assistant.

## Plain text for LLMs

For tools that consume documentation as a single file:

* [`/llms.txt`](https://ebrcapi.docs.eximfiles.io/llms.txt) is a compact index of every page.
* [`/llms-full.txt`](https://ebrcapi.docs.eximfiles.io/llms-full.txt) is the entire documentation as one plain-text file, ready to paste into a model's context.

## Call the API with AI

Reading the docs is one half; the other is making requests.

1. Open the [API Reference](/api-reference/introduction). Every endpoint has a live playground and copy-paste cURL, generated from the OpenAPI spec.
2. Add your `x-api-key` (see [Authentication](/authentication)) and the base URL for your environment (see [Environments](/environments)).
3. An MCP-connected assistant can read an endpoint page and generate correct request code for your language, because the request and response shapes come straight from the spec.

Your API key is yours: it goes in the `x-api-key` header over HTTPS on requests **you** make. It is never part of the docs, the MCP server, or any example on this site.

## When the base URLs change

The system is built so a future move is a swap, not a rebuild:

* The **API base URLs** always come from [Environments](/environments); they are driven by the `servers` in the OpenAPI spec, so a change updates the reference, the playground, and every cURL example at once. Take the current base URL from that page rather than hardcoding it.
* The **docs MCP URL** follows this documentation site; if the docs move to a new domain, use the MCP URL shown on this page at that time.

## Related

* [Quick Start](/quick-start) if you are integrating by hand.
* [Authentication](/authentication) for the API key header and rate limits.
* [API Reference](/api-reference/introduction) for conventions and the response envelope.