DataHub Metadata Platform

DataHub is the leading open source metadata platform evolved into a context platform for AI agents, trusted by an open source community of over 3,000 organizations and 15,000 members. Originally developed at LinkedIn and open sourced in 2020, the platform provides a unified context graph that connects metric definitions, data products, domain ownership, runbooks, and standard operating procedures, curated once and served to every agent, human, and dashboard that needs them. The platform delivers trusted lineage from pipeline to prompt, enabling teams to see where data comes from, how it is used, and how changes cascade through the system, for both humans debugging issues and AI agents answering queries. Data quality capabilities include freshness, volume, and column-level checks delivered to agents at query time rather than after the fact. The platform integrates with major data infrastructure including Microsoft Fabric, Databricks, Snowflake, Google Dataplex, BigQuery, Redshift, Looker, and dbt for structured data; Slack, Google Drive, Notion, and Confluence for unstructured data; Salesforce, Workday, and HubSpot for business applications. Enterprise customers include Saxo Bank, Janestreet, Robinhood, Credit Suisse, Verizon, and Thoughtspot. DataHub supports the Model Context Protocol (MCP), enabling AI coding agents like Claude and Cursor to leverage curated metadata context for more accurate data-driven applications.