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 of AI agents and human users lacking the business meaning, ownership, and trust signals needed to use raw data safely and effectively. It prevents AI from relying solely on raw database schemas by providing governed metadata, organizational memory, and semantic relationships.

How it works

OpenMetadata builds a schema-first metadata knowledge graph by collecting data from over 130 connectors. It normalizes this technical metadata and connects it to business semantics (glossaries, metrics), governance (policies, contracts), and "memory nuggets" (captured decisions and assumptions). This graph is then made actionable via an MCP server, semantic search, and AI SDKs, allowing LLMs and agents to query the context of their data estate.

Who it’s for

Data engineers, AI architects, and developers building AI agents or data discovery tools who need to ensure their AI systems have governed, semantic context and a way to preserve tribal knowledge.

Highlights

  • Context Graph: Connects technical metadata, quality signals, lineage, and business semantics into a single knowledge graph.
  • Organizational Memory: Captures conversations, AI threads, and remediation notes as governed entities to preserve tribal knowledge.
  • MCP Server: Provides a Model Context Protocol server that allows AI assistants to search metadata, inspect lineage, and update descriptions using natural language.
  • Broad Connectivity: Includes 130+ connectors and supports open standards like OpenLineage, DCAT, and ODCS for interoperability.
  • Semantic Search: Enables finding assets by business meaning rather than just exact keyword matches.

Related

  • Project
    datahub-project/datahub

    An open-source metadata management platform that turns enterprise data into trusted context for humans and AI agents through discovery, governance, and observability.

  • Project
    aws/context-ontology-accelerator

    A semantic context layer for AWS that combines knowledge graphs and formal ontologies to provide AI agents with validated, business-aware context.

  • Project
    alibaba/UnifiedModel

    UModel (Unified Model) is an open‑source, vendor‑neutral runtime that lets enterprises describe data, services, and infrastructure as a semantic graph. It provides model packs, a workspace‑scoped graph, unified query APIs, a web UI, CLI, and SDKs, and includes AgentGateway/MCP so LLM agents can discover, query, and act on the graph for tasks like root‑cause analysis.

  • 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
    thecodacus/understory

    understory is an open‑source, markdown‑based knowledge‑graph memory layer for LLM agents. It provides a deterministic OKF‑compliant bundle, a server exposing tool‑calling endpoints (memory_query, memory_add, etc.), a React UI with a force‑directed graph, CLI utilities, and multi‑profile isolation. Works with any OpenAI‑ or Anthropic‑compatible model (including local llama.cpp) and runs via Docker or from source.