MCP

dat.ai exposes a remote Model Context Protocol (MCP) server over streamable HTTP. Compatible clients can call the same capabilities as the HTTP API — chat, browsing, and transcription — as MCP tools.

Field	Value
Endpoint	`https://llm.dat.ai/mcp`
Transport	Streamable HTTP (not stdio)
Auth	`Authorization: Bearer YOUR_API_KEY`

Create a key in the dashboard under API Keys. The MCP server is a thin proxy: every tool maps to an existing HTTP endpoint.

Setup

Create an API key

In the dashboard open API Keys, create a key, and copy it. Treat it as a secret — do not commit configs that contain real keys.

Add the server to your client

Use the tab for your client below. Replace YOUR_API_KEY with the key from step 1.

Verify with `ping`

After the client connects, call the ping tool. A healthy response looks like:


{
  "ok": true,
  "has_api_key": true
}

If has_api_key is false, the Authorization header did not reach the server — fix the client config and reconnect.

Client configuration

Cursor

Cursor MCP settings (mcpServers):


{
  "mcpServers": {
    "llm-dat-ai": {
      "url": "https://llm.dat.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Claude Code

CLI:


claude mcp add --transport http llm-dat-ai https://llm.dat.ai/mcp \
  --header "Authorization: Bearer YOUR_API_KEY"

Or in .mcp.json / Claude settings — type is required when using a URL:


{
  "mcpServers": {
    "llm-dat-ai": {
      "type": "http",
      "url": "https://llm.dat.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

VS Code

Add to .vscode/mcp.json (or user MCP settings):


{
  "servers": {
    "llm-dat-ai": {
      "type": "http",
      "url": "https://llm.dat.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Windsurf

Windsurf uses serverUrl (not url) — typically ~/.codeium/windsurf/mcp_config.json:


{
  "mcpServers": {
    "llm-dat-ai": {
      "serverUrl": "https://llm.dat.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Then refresh MCP servers in Cascade → Manage MCPs.

Claude Desktop

Claude Desktop’s config file only speaks stdio. Bridge the remote HTTP server with mcp-remote:


{
  "mcpServers": {
    "llm-dat-ai": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://llm.dat.ai/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}"
      ],
      "env": {
        "AUTH_HEADER": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Requires Node.js. Keep the token in env (not inline in args) so spaces in Bearer … do not break the launch command.

Any other client that supports remote streamable HTTP MCP can use the same endpoint and Bearer header. Stdio-only clients need a bridge such as mcp-remote.

Available tools

Tools mirror the public HTTP API. Your client shows each tool’s input schema; details below match the server implementation.

Tool	Description
`ping`	Health check; returns `ok` and `has_api_key`. Does not start a task.
`chat_complete`	OpenAI-compatible chat → `POST /v1/chat/completions`. Optional `stream`; final result always includes the full assistant message.
`browsing_run`	Run browsing and wait → `POST …/browsing/sync`.
`browsing_start`	Start browsing async → `POST …/browsing/async`; returns `task_id`.
`browsing_status`	Poll browsing by `task_id` → `GET …/browsing/status`.
`browsing_screenshot`	Download a task screenshot as MCP image content. Filename comes from the task payload.
`browsing_list`	List recent browsing tasks (`page` / `limit`).
`whisper_transcribe_run`	Transcribe and wait → `POST …/transcribe/sync`. Audio must be base64.
`whisper_transcribe_start`	Start transcription async → `POST …/transcribe/async`; returns `task_id`.
`whisper_transcribe_status`	Poll transcription by `task_id` → `GET …/transcribe/status`.

For full HTTP request/response shapes, see Inference, Browsing, and Transcribing.

Tool input examples

Chat

chat_complete — required: model, messages.


{
  "model": "qwen3:1.7b",
  "messages": [
    { "role": "user", "content": "Hello!" }
  ],
  "temperature": 0.7
}

Optional: top_p, max_tokens, seed, stop, stream, and datai (tools.net / tools.fs / tools.webview, tps: fast | med | slow). See built-in tools.

Browsing

browsing_run / browsing_start — provide task (natural language) and/or script (custom Python, approved accounts only).


{
  "task": "Open https://example.com and return the page title.",
  "timeout_sec": 300,
  "fanout": 1,
  "screenshots": {
    "mode": "final",
    "full_page": true
  },
  "filter": {
    "country_iso": "US"
  }
}

timeout_sec is in seconds (max 3 hours = 10800). After browsing_start, poll with browsing_status using the returned task_id.

Transcription

whisper_transcribe_run / whisper_transcribe_start — required: audio_base64 (MCP inputs are JSON, so raw files cannot be uploaded).


{
  "audio_base64": "BASE64_AUDIO_CONTENT",
  "mime_type": "audio/mpeg"
}

Optional: keep_alive_ms, fanout. After whisper_transcribe_start, poll with whisper_transcribe_status.

Usage patterns

Goal	Approach
Check connectivity	`ping` — confirm `has_api_key: true`
Short chat	`chat_complete`
Browser task with immediate result	`browsing_run`
Long browser task	`browsing_start` → poll `browsing_status` → optional `browsing_screenshot`
Short audio	`whisper_transcribe_run`
Long audio / client timeouts	`whisper_transcribe_start` → poll `whisper_transcribe_status`

Browsing and transcription can run for many minutes. Prefer async tools when the client may time out waiting on a single tool call. The public proxy allows long MCP sessions (on the order of ~12 minutes).

Errors

Failures are returned as MCP tool errors. Common causes:

Cause	Fix
Missing or invalid API key	Set `Authorization: Bearer YOUR_API_KEY` with an active key from the dashboard
`has_api_key: false` on `ping`	Header not forwarded — check client config / `mcp-remote` bridge
Missing required field	Check the tool schema in your client (`model`/`messages`, `task`/`script`, `audio_base64`, `task_id`, …)
Task timeout / no nodes	Retry, raise `timeout_sec`, adjust filters, or use async tools
Invalid base64 audio	Re-encode the file as base64
Insufficient credits	Top up in the dashboard billing

MCP