# mcp-server-datahub A [Model Context Protocol](https://modelcontextprotocol.io/) server implementation for [DataHub](https://datahubproject.io/). This enables AI agents to query DataHub for metadata and context about your data ecosystem. Supports both DataHub Core and DataHub Cloud. ## Features - Searching across all entity types and using arbitrary filters - Fetching metadata for any entity - Traversing the lineage graph, both upstream and downstream - Listing SQL queries associated with a dataset ## Demo Check out the [demo video](https://youtu.be/VXRvHIZ3Eww?t=1878), done in collaboration with the team at Block. ## Usage 1. Install [`uv`](https://github.com/astral-sh/uv) ```bash # On macOS and Linux. curl -LsSf https://astral.sh/uv/install.sh | sh ``` 2. Locate your authentication details For authentication, you'll need the following: - The URL of your DataHub instance e.g. `https://tenant.acryl.io/gms` - A [personal access token](https://datahubproject.io/docs/authentication/personal-access-tokens/)

Alternative: Using ~/.datahubenv for authentication

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

3. Configure your MCP client. See below - this will vary depending on your agent. ### Claude Desktop Run `which uvx` to find the full path to the `uvx` command. In your `claude_desktop_config.json` file, add the following: ```json { "mcpServers": { "datahub": { "command": "", # e.g. /Users/hsheth/.local/bin/uvx "args": ["mcp-server-datahub"], "env": { "DATAHUB_GMS_URL": "", "DATAHUB_GMS_TOKEN": "" } } } } ``` ### Cursor In `.cursor/mcp.json`, add the following: ```json { "mcpServers": { "datahub": { "command": "uvx", "args": ["mcp-server-datahub"], "env": { "DATAHUB_GMS_URL": "", "DATAHUB_GMS_TOKEN": "" } } } } ``` ### Other MCP Clients ```yaml command: uvx args: - mcp-server-datahub env: DATAHUB_GMS_URL: DATAHUB_GMS_TOKEN: ``` ### Troubleshooting #### `spawn uvx ENOENT` The full stack trace might look like this: ``` 2025-04-08T19:58:16.593Z [datahub] [error] spawn uvx ENOENT {"stack":"Error: spawn uvx ENOENT\n at ChildProcess._handle.onexit (node:internal/child_process:285:19)\n at onErrorNT (node:internal/child_process:483:16)\n at process.processTicksAndRejections (node:internal/process/task_queues:82:21)"} ``` Solution: Replace the `uvx` bit of the command with the output of `which uvx`. ## Developing See [DEVELOPING.md](DEVELOPING.md).