# OpenMetadata [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/open-metadata/OpenMetadata  
**GitHub Stars:** 15552  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/openmetadata

## Description
Official OpenMetadata MCP: governed context and business semantics for AI assistants and agents.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "openmetadata": {
    "command": "uvx",
    "args": ["data-ai-sdk"]
  }
}
```

## Documentation

## What OpenMetadata does

OpenMetadata MCP gives AI assistants and agents access to the context stored in OpenMetadata’s metadata graph. That context spans databases, schemas, tables, columns, dashboards, charts, pipelines, APIs, search indexes, ML models, storage assets, topics, and other data-related resources. It also includes descriptions, joins, sample queries, owners, teams, domains, usage, and data products.

The graph adds the information needed to interpret and govern those assets. Agents can work with business terms, glossaries, metrics, KPIs, classifications, tags, policies, contracts, lifecycle states, certifications, and review workflows. Quality and trust signals include tests, suites, profiling results, freshness and volume checks, incidents, alerts, and observability history. Lineage connects upstream and downstream assets, including tables, columns, dashboards, pipelines, metrics, ML models, APIs, topics, and OpenLineage events.

OpenMetadata also preserves organizational knowledge. Conversations, AI threads, decisions, assumptions, runbooks, remediation notes, incident learnings, and reusable memory nuggets can be associated with assets, teams, users, data products, and agent workflows.

## How it works

OpenMetadata uses a schema-first metadata graph. Its platform collects metadata from across a data ecosystem through more than 130 connectors, ingestion APIs, events, and SDKs. The collected information is normalized through open schemas and standards, then connected into relationships between assets, people, policies, quality checks, lineage, semantics, contracts, and memories.

MCP is one of the ways OpenMetadata activates this graph for AI systems. Other access and activation paths include semantic search, APIs, SDKs, events, webhooks, metadata applications, and AI workflows. The graph structure lets an agent follow relationships such as a dashboard depending on a table, a column carrying a classification, a pipeline producing a table, or a data contract applying to an asset.

Memory is represented as part of the same governed context rather than as a separate conversational store. A decision or conversation can be retained as a reusable memory and linked to the asset, metric, data product, or workflow it concerns.

## Tools and capabilities

- Exposes governed metadata context to AI assistants and agents through MCP.
- Provides technical metadata for data assets and their structures.
- Connects upstream and downstream lineage, including column-level relationships.
- Surfaces quality, freshness, certification, testing, incidents, and trust signals.
- Relates technical assets to glossaries, metrics, classifications, policies, domains, and data products.
- Preserves conversations, decisions, assumptions, runbooks, and remediation knowledge as reusable memory.
- Represents ownership, stewardship, teams, contracts, roles, and governance context.
- Supports interoperability through standards including OpenLineage, DCAT, DPROD, PROV-O, RDF/OWL, JSON-LD, SHACL, JSON Schema, and ODCS.

The result is context for questions such as what a metric means, which datasets feed a dashboard, who owns a data product, whether a dataset is trusted, what downstream systems may be affected by a change, and which policies or business concepts apply.

_Full upstream README: https://allmcps.com/mcp/openmetadata/readme_