DataHub

Provides an MCP server and library to connect AI assistants to DataHub metadata catalogs for searching datasets, tracing lineage, and accessing glossary terms.

GitHub Stars

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "txn2-mcp-datahub": {
      "command": "/opt/homebrew/bin/mcp-datahub",
      "args": [],
      "env": {
        "DATAHUB_URL": "https://datahub.example.com",
        "DATAHUB_DEBUG": "false",
        "DATAHUB_TOKEN": "your_token",
        "DATAHUB_TIMEOUT": "30",
        "DATAHUB_MAX_LIMIT": "100",
        "DATAHUB_DEFAULT_LIMIT": "10",
        "DATAHUB_CONNECTION_NAME": "prod",
        "DATAHUB_ADDITIONAL_SERVERS": "{\"staging\":{\"url\":\"https://staging.datahub.example.com/api/graphql\",\"token\":\"staging-token\"}}"
      }
    }
  }
}

You run an MCP server that connects AI assistants to DataHub metadata catalogs, enabling you to search datasets, explore schemas, trace lineage, and access glossary terms and domains from your AI workflows.

How to use

You use an MCP client to connect to the DataHub MCP server and interact with DataHub metadata through AI assistants. Start by ensuring the server is running on your machine or accessible from your network, then configure your MCP client to point at the server. With the integration in place, you can perform searches for datasets, inspect dataset schemas, view lineage relationships, and retrieve glossary terms and domain information to enrich your conversations and automated workflows.

How to install

Prerequisites you should have before installing include a functioning Go toolchain or the ability to install apps via your package manager. If you plan to install and run as a local MCP server, follow the steps below to use the provided CLI bundle and environment configuration.

Install the MCP server binary via the included package or from a local bundle.
Prepare the environment variables required to connect to DataHub and secure authentication tokens.
Launch the MCP server using the designated startup command and verify it is reachable from your AI assistants.

Configuration notes

The server can be run as a local process with a dedicated command and environment variables to connect to your DataHub instance. The following configuration shows the required command and environment variables for a typical setup.

{
  "mcpServers": {
    "datahub": {
      "command": "/opt/homebrew/bin/mcp-datahub",
      "env": {
        "DATAHUB_URL": "https://datahub.example.com",
        "DATAHUB_TOKEN": "your_token"
      }
    }
  }
}

Connecting multiple DataHub instances

If you need to connect to more than one DataHub instance, configure each connection and let your MCP client discover available connections. This enables you to switch contexts between production, staging, and other environments from a single control plane.

Bidirectional context and middleware

You can enable bidirectional context injection so that query engines provide execution context back to DataHub tools. This enhances search results, entity details, and schema usage with execution mappings and sample queries. You can also add enterprise middleware for access control and audit logging to tailor and monitor tool usage.

Available tools

datahub_search

Search DataHub datasets, dashboards, and pipelines by query and type

datahub_get_entity

Retrieve entity metadata by URN including description, owners, tags, and domain

datahub_get_schema

Fetch dataset schema with field types and descriptions

datahub_get_lineage

Obtain upstream and downstream data lineage information

datahub_get_column_lineage

Obtain column-level lineage mappings

datahub_get_queries

Get SQL queries associated with a dataset

datahub_get_glossary_term

Get glossary term definitions and properties

datahub_list_tags

List available tags in the catalog

datahub_list_domains

List data domains

datahub_list_data_products

List data products

datahub_get_data_product

Get data product details including owners and properties

datahub_list_connections

List configured DataHub server connections (multi-server mode)

Built by

VeilStrat

AI signals for GTM teams