Skip to content

txn2/mcp-datahub

An MCP server that connects AI assistants to DataHub metadata catalogs. Search datasets, explore schemas, understand lineage, and access business context like glossary terms and domains.

Unlike other MCP servers, mcp-datahub is designed as a composable Go library. Import it into your own MCP server to add DataHub capabilities with custom authentication, tenant isolation, and audit logging. The standalone server works out of the box; the library lets you build exactly what your organization needs.

Get Started View on GitHub

Two Ways to Use

  • Use the Server

    Connect Claude, Cursor, or any MCP client to DataHub with secure defaults.

    • Search across all assets
    • Schema exploration
    • Lineage visualization

    Install in 5 minutes

  • Build Custom MCP

    Import the Go library for enterprise servers with auth, tenancy, and compliance.

    • OAuth, API keys, SSO
    • Row-level tenant isolation
    • SOC2 / HIPAA audit logs

    View library docs

Core Capabilities

  • Composable Architecture

    Import as a Go library to build custom MCP servers with authentication, tenant isolation, and audit logging without forking.

    Library docs

  • Metadata Catalog

    Access business descriptions, ownership, tags, domains, glossary terms, and data quality information from your DataHub instance.

    Tools reference

  • Lineage Exploration

    Understand upstream and downstream dependencies for datasets, dashboards, and pipelines with configurable depth.

    Configuration

  • Secure Defaults

    Token-based authentication, read-only operations, and SLSA Level 3 provenance for production deployments.

    Security reference

Available Tools

Tool Description
datahub_search Search across all DataHub assets
datahub_get_entity Get entity metadata by URN
datahub_get_schema Get dataset schema with field details
datahub_get_lineage Explore upstream/downstream dependencies
datahub_get_column_lineage Get fine-grained column-level lineage
datahub_get_queries Get SQL queries associated with a dataset
datahub_get_glossary_term Get term definition and relationships
datahub_list_tags List available tags in the catalog
datahub_list_domains List organizational domains
datahub_list_data_products List data products in catalog
datahub_get_data_product Get data product details and assets
datahub_list_connections List configured server connections

Pair mcp-datahub with txn2/mcp-trino for a complete data stack. mcp-trino queries your Trino data warehouse and can use DataHub as a semantic layer to enrich query results with business context.

Works With

Claude Desktop · Claude Code · Cursor · Windsurf · Any MCP Client