Open-source CLI for dbt-native data observability that reads Elementary dbt package results to generate reports and send Slack and Teams alerts.
- Stars
- 15,522
- Forks
- 2,466
- Open issues
- 819
- Last commit
- 6 hours ago
What is it?
OpenMetadata is an open platform that collects metadata from warehouses, lakes, BI tools, pipelines and other systems through 130+ connectors, then links it into a single knowledge graph. The graph covers table and column-level lineage, owners, glossary terms, classifications, data quality tests, data contracts and data products, so teams can find, understand and trust their data.
The project also includes an MCP server that lets compatible AI assistants search metadata, run semantic search, inspect lineage, create glossary terms and tests, and update descriptions or owners. It offers memory entities for storing decisions, assumptions and remediation notes attached to assets. Python, TypeScript and Java SDKs and REST APIs are provided for ingestion and automation.
The platform is built on open schemas, with support for ODCS 3.1 data contracts, OpenLineage events and DCAT/DPROD and PROV-O semantic models. It is written primarily in TypeScript and released under Apache-2.0.
Key features
- 01Connects to 130+ data sources including warehouses, BI tools and pipelines
- 02Provides table and column-level lineage with impact analysis
- 03Includes data quality tests, profiling and freshness checks
- 04Offers glossaries, classifications, domains and data contracts for governance
- 05Exposes an MCP server and SDKs so AI assistants can query and update metadata
Use cases
- Finding and understanding datasets across a company's data warehouse and BI tools
- Assessing what downstream dashboards or models break when a column changes
- Giving AI assistants governed access to business definitions and data ownership
- Tracking data quality issues and recording how they were fixed
Quick start
- 1
Install the Python SDK for reading and writing metadata
pip install "openmetadata-ingestion" - 2
Connect to a running OpenMetadata server with your token and read a table
- 3
Optionally install the AI SDK to give an LLM or agent governed access via MCP
pip install data-ai-sdk
Things to know
- The SDK examples need a running OpenMetadata server and an access token.
- The SDK version must match the server version.
Topics
FAQ
- What is OpenMetadata?
- Open-source metadata platform that unifies data catalog, lineage, quality, governance and an MCP server to give humans and AI agents context about data. OpenMetadata is an open platform that collects metadata from warehouses, lakes, BI tools, pipelines and other systems through 130+ connectors, then links it into a single knowledge graph. The graph covers table and column-level lineage, owners, glossary terms, classifications, data quality tests, data contracts and data products, so teams can find, understand and trust their data.
- Is OpenMetadata free to use? What license does it have?
- OpenMetadata is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is OpenMetadata still maintained?
- Actively maintained. The last commit to OpenMetadata was on Oct 20, 2026.
- What can I use OpenMetadata for?
- Finding and understanding datasets across a company's data warehouse and BI tools Assessing what downstream dashboards or models break when a column changes Giving AI assistants governed access to business definitions and data ownership Tracking data quality issues and recording how they were fixed
- How do I install OpenMetadata?
- Install the Python SDK for reading and writing metadata: pip install "openmetadata-ingestion" Connect to a running OpenMetadata server with your token and read a table Optionally install the AI SDK to give an LLM or agent governed access via MCP: pip install data-ai-sdk
- Where can I get OpenMetadata?
- The source code and installation instructions are on GitHub: https://github.com/open-metadata/OpenMetadata
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Local-first AI memory system that stores conversation history verbatim and retrieves it with semantic search, exposed as an MCP server and CLI.
Local AI search engine that auto-indexes any folder (code, docs, logs, PDFs) and speaks the Elasticsearch API, built in Rust.
A private, self-hosted search engine that indexes the web pages you visit and the files you keep, searchable from web, terminal, or AI assistants via MCP.
Desktop and self-hosted app that uses AI to organize your GitHub stars, with semantic search, repo Q&A, MCP access and release tracking.
Agent memory system that stores, recalls and reflects on memories so AI agents learn over time instead of only replaying chat history.