Skip to main content
Version: Next

DataHub MCP Server

The DataHub MCP Server implements the Model Context Protocol (MCP), which standardizes how applications provide context to LLMs and AI agents. This enables AI agents to query DataHub metadata and use it to find relevant assets, traverse lineage, and more.

Want to learn more about the motivation, architecture, and advanced use cases? Check out our deep dive blog post.

There's two ways to use the MCP server, which vary in setup required but offer the same capabilities.

Capabilities​

Search for Data
Find the right data to use for your projects & analysis by asking questions in plain English - skip the tribal knowledge.

Dive Deeper
Separate the signal from noise with rich context about your data, including usage, ownership, documentation, tags, and quality.

Lineage & Impact Analysis
Understand the impact of upcoming changes to tables, reports, and dashboards using DataHub’s end-to-end lineage graph.

Query Analysis & Authoring
Understand how your mission-critical data is typically queried, or build custom queries for your tables.

Works Where You Work
Seamlessly integrates with AI-native tools like Cursor, Windsurf, Claude Desktop, and OpenAI to supercharge your workflows.

Managed MCP Server Usage​

For folks on DataHub Cloud v0.3.12+, you can use our hosted MCP server endpoint.

note

The managed MCP server endpoint is only available with DataHub Cloud v0.3.12+. For DataHub Core and older versions of DataHub Cloud, you'll need to self-host the MCP server.

Prerequisites​

To connect to the MCP server, you'll need the following:

Your hosted MCP server URL is https://<tenant>.acryl.io/integrations/ai/mcp/?token=<token>.

Configure​

Claude Desktop
  1. Open your claude_desktop_config.json file. You can find it by navigating to Claude Desktop -> Settings -> Developer -> Edit Config.
  2. Update the file to include the following content. Be sure to replace <tenant> and <token> with your own values.
{
  "mcpServers": {
    "datahub-cloud": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://<tenant>.acryl.io/integrations/ai/mcp/?token=<token>"
      ]
    }
  }
}
Cursor
  1. Navigate to Cursor -> Settings -> Cursor Settings -> MCP -> add a new MCP server
  2. Enter the following into the file:
{
  "mcpServers": {
    "datahub-cloud": {
      "url": "https://<tenant>.acryl.io/integrations/ai/mcp/?token=<token>"
    }
  }
}
  1. Once you've saved the file, confirm that the MCP settings page shows a green dot and a couple tools associated with the server.
Other

Most AI tools support remote MCP servers. For those, you'll typically need to:

  • Provide the hosted MCP server URL: https://<tenant>.acryl.io/integrations/ai/mcp/?token=<token>
  • Ensure that the authentication mode is not set to "OAuth" (if applicable)

For AI tools that don't yet support remote MCP servers, you can use the mcp-remote tool to connect to the MCP server.

  • Command: npx
  • Args: -y mcp-remote https://<tenant>.acryl.io/integrations/ai/mcp/?token=<token>

Self-Hosted MCP Server Usage​

You can run the open-source MCP server locally.

Prerequisites​

  1. Install uv

    # On macOS and Linux.
    curl -LsSf https://astral.sh/uv/install.sh | sh

  2. For authentication, you'll need the following:

    • The URL of your DataHub instance's GMS endpoint; e.g. http://<localhost>:8080 or https://<tenant>.acryl.io
    • A personal access token

Configure​

Claude Desktop

  1. Run which uvx to find the full path to the uvx command.

  2. Open your claude_desktop_config.json file. You can find it by navigating to Claude Desktop -> Settings -> Developer -> Edit Config.

  3. Update the file to include the following content. Be sure to replace <tenant> and <token> with your own values.

{
  "mcpServers": {
    "datahub": {
      "command": "<full-path-to-uvx>",  // e.g. /Users/hsheth/.local/bin/uvx
      "args": ["mcp-server-datahub"],
      "env": {
        "DATAHUB_GMS_URL": "<your-datahub-url>",
        "DATAHUB_GMS_TOKEN": "<your-datahub-token>"
      }
    }
  }
}
Cursor
  1. Navigate to Cursor -> Settings -> Cursor Settings -> MCP -> add a new MCP server
  2. Enter the following into the file:
{
  "mcpServers": {
    "datahub": {
      "command": "uvx",
      "args": ["mcp-server-datahub"],
      "env": {
        "DATAHUB_GMS_URL": "<your-datahub-url>",
        "DATAHUB_GMS_TOKEN": "<your-datahub-token>"
      }
    }
  }
}
  1. Once you've saved the file, confirm that the MCP settings page shows a green dot and a couple tools associated with the server.
Other

For other AI tools, you'll typically need to provide the following configuration:

  • Command: uvx
  • Args: mcp-server-datahub
  • Env:
    • DATAHUB_GMS_URL: <your-datahub-url>
    • DATAHUB_GMS_TOKEN: <your-datahub-token>