open-metadata/OpenMetadata

The Open Context Layer for Data and AI , OpenMetadata is the open platform for building trusted data context and business semantics for humans, AI assistants, and agents.

What it solves

OpenMetadata provides a unified "context layer" for enterprise data, solving the problem where AI agents and human users lack the business meaning, ownership, and trust signals needed to use raw data safely and effectively. It prevents AI from having to rely solely on raw database connectors by providing governed metadata, organizational memory, and semantic relationships.

How it works

OpenMetadata builds a schema-first metadata graph that connects technical metadata (schemas, tables) with operational signals (quality, freshness), business semantics (glossaries, KPIs), and organizational memory (conversations, decisions). It uses over 130 connectors to collect data from across the ecosystem and normalizes it using open standards (like DCAT, PROV-O, and OpenLineage).

This context is then made actionable via an MCP server, semantic search, and dedicated AI SDKs, allowing LLMs and agents to query the metadata graph to understand data lineage, check for certifications, or retrieve historical decisions about a specific asset.

Who it’s for

  • AI Engineers building trusted AI assistants or agents that need to reason over enterprise data.
  • Data Engineers and Architects managing data catalogs, lineage, and data contracts.
  • Data Governance Officers ensuring policy-aware context and data quality across the organization.
  • Coding Agents that need to understand schemas and business definitions before generating SQL or dbt models.

Highlights

  • Organizational Memory: Captures conversations, AI threads, and remediation notes as "memory nuggets" tied to specific data assets.
  • MCP Server: Enables MCP-compatible AI assistants to interact with the metadata graph using natural language.
  • Extensive Connectivity: 130+ connectors for warehouses, lakes, BI tools, and ML platforms.
  • Semantic Search: Allows users and AI to find assets based on meaning rather than just keywords.
  • Open Standards: Built on interoperable standards including ODCS 3.1 for data contracts and JSON-LD for knowledge graphs.

Related

  • Project
    aws/context-ontology-accelerator

    An AWS‑hosted open‑source platform that scans data sources, builds a unified ontology/knowledge graph, and serves context to AI agents via SPARQL and a custom MCP interface. Includes Smithy‑defined APIs, CDK‑provisioned infrastructure, reasoning engines, and a React admin UI. Licensed Apache 2.0; read‑only mirror.

  • Project
    pipeshub-ai/pipeshub-ai

    An open-source enterprise AI context layer that connects organizational knowledge to power explainable RAG, AI agents, and search with built-in permission controls.

  • Project
    NateBJones-Projects/OB1

    An infrastructure layer that provides a unified, persistent memory database with vector search, allowing multiple AI tools and agents to share a single, consistent context of the user.

  • Dispatch
    Inside OpenAI's In-House Data Agent

    OpenAI has developed a bespoke internal AI data agent that enables employees to perform complex data analysis across 600 petabytes of data using natural language, utilizing a multi-layered context system and a self-learning reasoning loop.

  • Project
    marmotdata/marmot

    An open-source data catalog that provides a context layer for AI agents and teams to discover and track data assets like tables, APIs, and queues.