OpenMetadata

OpenMetadata is an open-source platform for data discovery, data cataloguing, lineage and governance that now positions itself as an open context layer for AI. It brings together technical metadata, data-quality signals, column-level lineage, ownership, usage, policies, glossaries, classifications, metrics, data contracts and organisational memory, so that people and AI assistants can understand what data exists, what it means, who owns it and whether it can be trusted.

It suits data teams who need a catalogue and governance layer across warehouses, lakes, BI tools, pipelines and ML platforms, and AI teams who want agents to answer data questions with business context rather than raw table access. Metadata is collected from many systems through connectors, and the context is made usable through semantic search, APIs, SDKs, events and an MCP server, so AI assistants can query the catalogue directly. Memories can be captured and tied to data assets, teams and incidents, turning tribal knowledge into reusable context.

Getting started: follow the official quickstart to run OpenMetadata, connect your first data sources through the ingestion connectors, then enable the MCP server to let assistants access the catalogue.