MCP Server Datahub

🚀 mcp-server-datahub

A server implementation of the Model Context Protocol on DataHub. This enables AI agents to query DataHub for metadata and contextual information about your data ecosystem.

It supports both the core version of DataHub and the DataHub Cloud version.

🚀 Quick Start

This section provides a step - by - step guide on how to use the mcp - server - datahub to enable AI agents to query DataHub for metadata and contextual information.

✨ Features

• Search for all entity types with arbitrary filters.
• Retrieve metadata for any entity.
• Traverse the lineage graph, including upstream and downstream directions.
• List SQL queries associated with datasets.

🎥 Demo

Check out the demo video, a collaborative effort by the Block team.

📦 Installation

1. Install uv

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

2. Find your authentication information To authenticate, you need the following:

   • Your DataHub instance URL, e.g., https://tenant.acryl.io/gms
   • A personal access token
   Alternative: Authenticate using ~/.datahubenv

   You can also use the `~/.datahubenv` file to configure your authentication. The easiest way is to run `datahub init` and follow the prompts. ```bash uvx --from acryl - datahub datahub init ```

💻 Usage Examples

Basic Usage

3. Configure your MCP client. The following examples vary by proxy.

Claude Desktop

Run which uvx to find the full path of the uvx command. Add the following to your claude_desktop_config.json file:

{
  "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

Add the following to your .cursor/mcp.json file:

{
  "mcpServers": {
    "datahub": {
      "command": "uvx",
      "args": ["mcp - server - datahub"],
      "env": {
        "DATAHUB_GMS_URL": "<your - datahub - url>",
        "DATAHUB_GMS_TOKEN": "<your - datahub - token>"
      }
    }
  }
}

Other MCP Clients

command: uvx
args:
  - mcp - server - datahub
env:
  DATAHUB_GMS_URL: <your - datahub - url>
  DATAHUB_GMS_TOKEN: <your - datahub - token>

Advanced Usage

Troubleshooting spawn uvx ENOENT

The full error stack might look like this:

2025-04-08T19:43:07.657Z - error: Error: spawn uvx ENOENT
    at ChildProcess.spawn (node:internal/child_process.js:407:11)
    at Object.spawnSync (node:internal/child_process.js:472:11)
    at Spawn - syncSpawn (file:///Users/hsheth/.config/chromium/Default/extensions/gppongmhjkannfkpmsamajpmgidpojnlp/6.30.0_198/core/index.js:153:223)

Solution: Replace uvx in the command with the actual path obtained from the system, e.g., /Users/hsheth/.local/bin/uvx.

📚 Documentation

For more information, refer to the [development documentation](https://github.com/plangz/mcp - server - datahub/blob/main/DEVELOPMENT.md).