Documentation Index

Fetch the complete documentation index at: https://docs.dataddo.com/llms.txt

Use this file to discover all available pages before exploring further.

Dataddo MCP Server

  • 2 minute(s) read
Prev Next

🚧 Prototype

The Dataddo MCP server is an early prototype. Its capabilities are limited and will expand quickly; tool names and behavior may change without notice. Do not build production workflows on it yet.

The Dataddo MCP server lets AI assistants and agents work with Dataddo directly from a conversation. It is built on the Model Context Protocol (MCP), an open standard that AI clients such as Claude, ChatGPT, Cursor, and Gemini use to discover and call external tools.

Once connected, your AI assistant can call Dataddo tools on your behalf - for example, finding the right connector for a data source and pulling up its datasets and setup instructions, all in natural language.

Server Details
Endpoint URL https://headless.dataddo.com/mcp
Transport Streamable HTTP
Authentication None. The prototype is open and requires no token. The current tools read only Dataddo's public connector catalog, so no account data is exposed.
Rate limits Shares the Headless API quota of 10,000 calls per day. Monitor the X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Retry-After response headers.

Available Tools

The prototype currently exposes the following tools:

Tool Description
current-time Returns the current date and time in UTC. Accepts an optional format argument (a PHP date() format string, default Y-m-d H:i:s).
find_connectors Maps free text (a search query or support request, for example "my Facebook Ads stopped syncing") to matching Dataddo connector IDs, each with a confidence hint and short supporting evidence. Returns nothing when no connector applies.
get_connector Given a connector ID (from find_connectors), returns the connector's datasets and full documentation: overview, authentication methods, prerequisites, connectivity (for example SSH tunnel or IP whitelisting), and setup steps.

A typical flow is to call find_connectors to identify the right connector from a plain-language description, then get_connector to retrieve its datasets and setup guide.

What's Coming

We are actively expanding the server. Planned capabilities include:

  • Pipeline management - create and inspect sources, destinations, and flows, mirroring what the Headless API offers today.
  • Data access - query the data Dataddo has extracted for you and get answers directly in your AI client.
  • Monitoring and troubleshooting - check flow health, review extraction errors, and diagnose failing pipelines.

Watch the Release Notes for updates.

Connect Your Client

Setup differs slightly per AI client. Follow the guide for yours:

Any other MCP-compatible client that supports the Streamable HTTP transport can connect using the endpoint URL above.

Security Considerations

  • No token is required today, and the current tools only read Dataddo's public connector catalog, so connecting the server does not expose your account data.
  • Account-touching tools (creating and inspecting sources, destinations, and flows) are on the roadmap and will introduce authentication. When they arrive, review any changes an assistant makes on your behalf, especially while the server is in the prototype stage.
  • AI assistants can misinterpret instructions. Treat their output as suggestions to verify, not as authoritative.

Troubleshooting

  • The client cannot connect. Verify the URL is exactly https://headless.dataddo.com/mcp and that your client supports the Streamable HTTP transport (SSE-only clients are not supported).
  • Rate limit errors. You have exhausted the daily quota shared with the Headless API. The X-RateLimit-Retry-After header indicates when the limit resets.

Need help? Contact us at support@dataddo.com.