Use the DataHub MCP server

Make your data ecosystem visible to AI agents with the DataHub MCP server, enabling natural language search, end-to-end lineage tracking, and context-aware SQL generation.

DataHub UI permissions:Generate Personal Access Tokens orManage All Access Tokens.
View the DataHub access control policies documentation

Prerequisites

• The uv Python package and project manager installed
• A DataHub personal access tokens (PAT)
• The DataHub GMS URL

Get the DataHub GMS URL

1. In the Aiven Console, go to your DataHub service.
2. In the Connection information section, copy the Application URL.
3. Add /api/gms to the end of the copied URL.

Configure the MCP server

To run the open-source MCP server locally, configure your AI assistant:

Claude Desktop

1. To find the full path to the uvx command, run:

   which uvx

2. Add the following content to your claude_desktop_config.json file:

   {
    "mcpServers": {
      "datahub": {
        "command": "FULL-PATH-TO-UVX",  // For example: /Users/hsheth/.local/bin/uvx
        "args": ["mcp-server-datahub@latest"],
        "env": {
          "DATAHUB_GMS_URL": "DATAHUB_APPLICATION_URL",
          "DATAHUB_GMS_TOKEN": "DATAHUB_PAT"
         }
       }
     }
   }

   Where:

   • FULL-PATH-TO-UVX is the full path to the uvx command you found.
   • DATAHUB_APPLICATION_URL is the DataHub GMS URL.
   • DATAHUB_PAT is the personal access token you created in the DataHub UI.

Cursor

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

     {
       "mcpServers": {
         "datahub": {
           "command": "uvx",
           "args": ["mcp-server-datahub@latest"],
           "env": {
             "DATAHUB_GMS_URL": "DATAHUB_APPLICATION_URL",
             "DATAHUB_GMS_TOKEN": "DATAHUB_PAT"
           }
         }
       }
     }

   Where:

   • DATAHUB_APPLICATION_URL is the DataHub GMS URL.
   • DATAHUB_PAT is the personal access token you created in the DataHub UI.

Other AI tools

The model context protocol (MCP) is an open standard supported by many clients. The most common configurations require:

• The DataHub GMS URL
• Your DataHub personal access token
• Command: uvx
• Args: mcp-server-datahub@latest

The following standard MCP JSON configuration format works with most clients:

"datahub": {
      "command": "/Users/USER_NAME/.local/bin/uvx",
      "args": [
        "mcp-server-datahub@latest"
      ],
      "env": {
        "DATAHUB_GMS_URL": "DATAHUB_APPLICATION_URL",
        "DATAHUB_GMS_TOKEN": "DATAHUB_PAT"
}

Where:

• DATAHUB_APPLICATION_URL is the DataHub GMS URL.
• DATAHUB_PAT is the personal access token you created in the DataHub UI.