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

# Overview

> Connect Claude, Cursor, and other AI agents to the Dataspike API over the Model Context Protocol — AML screening, verifications, KYB, KYT, email checks, and deepfake detection.

The **Dataspike MCP Server** exposes the Dataspike API to AI agents over the
[Model Context Protocol (MCP)](https://modelcontextprotocol.io). Once connected, an LLM host —
Claude Code, Cursor, and other MCP-capable clients — can run compliance tasks for
you in natural language, without writing any integration code.

Through the server an agent can:

* **AML screening** — screen people and organizations against sanctions, PEP, and watchlists, and pull past results.
* **Applicants** — look up applicant profiles and their latest AML reports.
* **Verifications** — inspect KYC verification sessions and their event timelines.
* **KYB** — poll business-verification results.
* **KYT** — retrieve crypto-wallet check reports.
* **Email verification** — fetch email-check results.
* **Deepfake detection** — run authenticity analysis on selfie and ID-document images.

See the full list on the [Tools](/mcp-server/tools) page.

***

## How it works

An MCP host connects to the server over HTTP and discovers the **tools** it exposes. The model
decides which tool to call, and with what arguments, based on your request — you just describe the
task. Results come back as structured data the agent can reason over and summarize.

The managed server is available at:

```
https://mcp.dataspike.io/mcp
```

***

## Authentication

The server authenticates to Dataspike with your API key, passed once in the client configuration as
a bearer token:

```
Authorization: Bearer <YOUR_API_KEY>
```

You configure this in your MCP client settings (see below) — there's no need to include the key in
your messages. Generate an API key in the [Dataspike Dashboard](https://dash.dataspike.io/api).

<Note>
  One tool, `list_aml_sources`, queries a public catalog endpoint and does not require an API key.
  Every other tool calls the authenticated Dataspike API.
</Note>

***

## Connect your client

<Tabs>
  <Tab title="Claude Desktop">
    Claude Desktop connects over OAuth as a **custom connector** — no manual header; you authorize once
    with your Dataspike API key. (Custom connectors are in beta — Free, Pro, Max, Team, and Enterprise.)

    <Steps>
      <Step title="Open connector settings">
        In Claude Desktop, go to **Customize → Connectors**.
      </Step>

      <Step title="Add a custom connector">
        Click **+**, then **Add custom connector**, and enter the URL:

        ```
        https://mcp.dataspike.io/mcp
        ```

        Leave **Advanced settings** empty — the server registers your client automatically. Click **Add**.
      </Step>

      <Step title="Authorize with your API key">
        Click **Connect**. A Dataspike sign-in page opens — paste your
        [Dataspike API key](https://dash.dataspike.io/api) and authorize. The connector shows as
        connected and the Dataspike tools become available in chat.
      </Step>
    </Steps>

    <Note>
      On **Team / Enterprise**, an owner first adds the connector under **Organization settings →
      Connectors → Add → Custom → Web**; members then **Connect** it from **Customize → Connectors**.
    </Note>
  </Tab>

  <Tab title="Claude Code">
    Add to `.claude/settings.json` in your project root:

    ```json theme={null}
    {
      "mcpServers": {
        "dataspike": {
          "type": "http",
          "url": "https://mcp.dataspike.io/mcp",
          "headers": {
            "Authorization": "Bearer <YOUR_API_KEY>"
          }
        }
      }
    }
    ```
  </Tab>

  <Tab title="Cursor">
    Open **Settings → MCP → Add server** and configure:

    * **URL:** `https://mcp.dataspike.io/mcp`
    * **Header:** `Authorization: Bearer <YOUR_API_KEY>`
  </Tab>

  <Tab title="Codex CLI">
    Edit `~/.codex/config.toml` (or `.codex/config.toml` in a trusted project). Codex reads the
    bearer token from an environment variable and sends it as `Authorization: Bearer <token>`:

    ```toml theme={null}
    [mcp_servers.dataspike]
    url = "https://mcp.dataspike.io/mcp"
    bearer_token_env_var = "DATASPIKE_API_KEY"
    ```

    Then export your key before launching Codex:

    ```bash theme={null}
    export DATASPIKE_API_KEY="<YOUR_API_KEY>"
    ```
  </Tab>
</Tabs>

***

## Example prompt

Once connected, just ask in plain language:

> Check John Smith for AML, country United Kingdom

The agent picks the right tool (`search_aml`), fills in the arguments, and reports the risk level
and matches.

***

## Next steps

<Columns cols={2}>
  <Card title="Tools" icon="wrench" href="/mcp-server/tools">
    Browse every tool the server exposes, grouped by domain, with parameters.
  </Card>

  <Card title="API Reference" icon="code" href="/api-reference">
    The underlying Dataspike endpoints, request/response schemas, and parameters.
  </Card>
</Columns>