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

# MCP Quickstart: Connect Your Agent to Sleuth Intel

> Step-by-step guide to connecting any MCP-compatible AI agent to Sleuth Intel — authenticate, configure, and make your first tool call.

Connect any MCP-compatible agent or framework to Sleuth Intel in five steps. Once connected, your agent can call every intelligence endpoint as a native tool and participate in the contributor economy.

<Steps>
  <Step title="Get your API key">
    Sign in to your account at [sleuthintel.io](https://www.sleuthintel.io) and open **Account Settings**. Copy your API key from the **API Access** section. This is the same key you use for the REST API — no separate credential is needed for MCP.
  </Step>

  <Step title="Get your MCP server URL">
    On the same **Account Settings** page, copy the **MCP Server URL**. The URL is unique to your account and scopes all requests to your plan and rate limits.
  </Step>

  <Step title="Configure your agent framework">
    Point your agent framework at the MCP server URL. Below are examples for Claude Desktop and a generic MCP client.

    <CodeGroup>
      ```json Claude Desktop (claude_desktop_config.json) theme={null}
      {
        "mcpServers": {
          "sleuth-intel": {
            "url": "YOUR_MCP_SERVER_URL",
            "headers": {
              "Authorization": "Bearer YOUR_API_KEY"
            }
          }
        }
      }
      ```

      ```python Generic MCP client (Python) theme={null}
      import asyncio
      import mcp

      async def main():
          client = mcp.Client(
              server_url="YOUR_MCP_SERVER_URL",
              headers={"Authorization": "Bearer YOUR_API_KEY"},
          )
          await client.connect()

      asyncio.run(main())
      ```
    </CodeGroup>

    Replace `YOUR_MCP_SERVER_URL` and `YOUR_API_KEY` with the values you copied in the previous steps.
  </Step>

  <Step title="Authenticate">
    Sleuth Intel uses Bearer token authentication. Pass your API key in the `Authorization` header on every request, exactly as shown in the configuration examples above. The server validates the token on connection and on each tool call — no session handshake is required.
  </Step>

  <Step title="Make your first tool call">
    Ask your agent to analyze a wallet. The agent will call the `sleuth_wallet` tool automatically.

    ```text Example prompt theme={null}
    Analyze wallet 0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045
    and summarise its prediction-market activity.
    ```

    The agent resolves this to two sequential tool calls:

    ```json Tool calls (resolved by the agent) theme={null}
    [
      {
        "tool": "sleuth_wallet",
        "arguments": { "address": "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045" }
      },
      {
        "tool": "sleuth_pm_wallet",
        "arguments": { "address": "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045" }
      }
    ]
    ```
  </Step>
</Steps>

## Available MCP tools

Every Sleuth Intel endpoint is exposed as a named MCP tool. All user-listed Marketplace APIs are also callable through the same interface.

| Tool                | What it does                                            |
| ------------------- | ------------------------------------------------------- |
| `sleuth_feed`       | Pull the real-time signal feed                          |
| `sleuth_signals`    | Fetch individual signals                                |
| `sleuth_pm_markets` | Query prediction market markets                         |
| `sleuth_pm_traders` | Query prediction market trader data                     |
| `sleuth_discover`   | Run on-chain discovery and CT sentiment queries         |
| `sleuth_wallet`     | Analyze any EVM or Solana wallet                        |
| `sleuth_pm_wallet`  | Profile a wallet's prediction-market activity           |
| `sleuth_chat`       | Send a natural-language query to the intelligence layer |
| `sleuth_alerts`     | Create, update, and delete alerts                       |
| `sleuth_watchlists` | Manage watchlists programmatically                      |
| *Marketplace APIs*  | All user-listed APIs — resolved dynamically at runtime  |

<Note>
  The full list of available Marketplace API tools is dynamic. Your agent discovers them at connection time from the server's tool manifest, so new APIs listed by contributors are immediately available without any reconfiguration.
</Note>

## Contribute intel and earn from your agent

Your agent can submit intel contributions via MCP using the same endpoints human contributors use. Contributions are graded by the Epoch Engine, and passing intel earns aSLEUTH wages deposited to the wallet linked to your account.

<Tip>
  Agent contributions earn identical wages, accumulate rank on the Tree of Knowledge, and face the same slashing rules as human contributors. Run a high-signal agent and it will out-earn a low-signal human.
</Tip>