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

# AI Integrations

> Use EQX with AI agents via MCP and llms.txt

EQX exposes its integrator API surface to AI agents and LLM-powered tools through two mechanisms: an **MCP server** and an **`llms.txt` manifest**.

## MCP Server

The EQX MCP (Model Context Protocol) server lets AI agents like Claude and Cursor call EQX API operations directly as tools.

**Endpoint:**

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

**Connect in Claude Desktop** (`claude_desktop_config.json`):

```json theme={null}
{
  "mcpServers": {
    "eqx": {
      "url": "https://docs.eqxpay.io/mcp",
      "headers": {
        "Authorization": "Bearer <your_token>"
      }
    }
  }
}
```

Once connected, Claude can call tools like `create_quote`, `commit_transaction`, `list_beneficiaries`, and `get_transaction_detail` directly from a conversation.

**Example prompt:**

> "Create a KES payout of 100 USDC to beneficiary ben\_01J5K2 using M-Pesa."

The MCP server maps this to `POST /api/quotes` followed by `POST /api/transactions/commit`.

***

## llms.txt

The EQX `llms.txt` file provides a machine-readable index of the integrator API surface — optimised for LLM context windows.

```
https://docs.eqxpay.io/llms.txt
```

Use it to prime an LLM with EQX API knowledge before asking it to generate integration code:

```bash theme={null}
curl https://docs.eqxpay.io/llms.txt | pbcopy
# Paste into your AI tool's context
```

The file is automatically updated whenever the OpenAPI spec changes via CI.