# OpenMetadata — Governed Data Context via MCP > OpenMetadata is an open-source metadata platform that builds a governed data graph (lineage, owners, quality). It ships an MCP server at `/mcp`. ## Install Merge the JSON below into your `.mcp.json`: ## Quick Use ```bash # Try the sandbox echo 'https://sandbox.open-metadata.org' # MCP endpoint (served by your OpenMetadata instance) echo 'https:///mcp' # MCP docs echo 'https://docs.open-metadata.org/latest/how-to-guides/mcp' ``` ## Intro OpenMetadata is an open-source metadata platform that builds a governed data graph (lineage, owners, quality). It ships an MCP server at `/mcp`. **Best for:** Data teams who want AI agents to query governed metadata safely **Works with:** OpenMetadata server, MCP-compatible clients, semantic search **Setup time:** 20-60 minutes ### Key facts (verified) - GitHub: 13901 stars · 2095 forks · pushed 2026-05-13. - License: Apache-2.0 · owner avatar + repo URL verified via GitHub API. - README-verified entrypoint: `echo 'Follow https://docs.open-metadata.org/latest/quick-start'`. ## Main - Install OpenMetadata using the project’s quick-start docs, then enable the MCP application and connect your MCP client to `/mcp`. - Use MCP as the governed interface: lineage, ownership, glossary terms, tags/classifications, and quality results become agent-queryable context. - Pair MCP with Semantic Search for natural-language discovery when users don’t know exact asset names. ### Source-backed notes - README states OpenMetadata includes an MCP server and links to MCP documentation. - README Quickstart section shows the MCP endpoint format: `https:///mcp`. - README lists MCP tools/workflows and calls out Semantic Search through MCP as a supported workflow. ### FAQ - **Do I need MCP to use OpenMetadata?**: No — MCP is an integration option; the platform also exposes APIs and UI for metadata work. - **What does MCP unlock?**: README frames MCP as a way for assistants/agents to interact with the metadata graph via natural language. - **What should I set up first?**: Start with ingestion + basic governance (owners, glossary, quality) so MCP answers have reliable context. ## Source & Thanks > Source: https://github.com/open-metadata/OpenMetadata > License: Apache-2.0 > GitHub stars: 13901 · forks: 2095 --- ## Quick Use ```bash # Try the sandbox echo 'https://sandbox.open-metadata.org' # MCP endpoint (served by your OpenMetadata instance) echo 'https:///mcp' # MCP docs echo 'https://docs.open-metadata.org/latest/how-to-guides/mcp' ``` ## Intro OpenMetadata 是开源元数据平台,把血缘、质量、负责人、语义等连成知识图谱。README 明确包含 MCP server,并给出面向 agent 的 `/mcp` 端点与文档入口。 **Best for:** 希望让 AI agent 在可治理前提下访问元数据/血缘/质量信息的数据团队 **Works with:** OpenMetadata 服务端;MCP 客户端;语义搜索 **Setup time:** 20-60 minutes ### Key facts (verified) - GitHub:13901 stars · 2095 forks;最近更新 2026-05-13。 - 许可证:Apache-2.0;作者头像与仓库链接均已通过 GitHub API 复核。 - README 中核对过的入口命令:`echo 'Follow https://docs.open-metadata.org/latest/quick-start'`。 ## Main - 先按 Quick-start 文档部署 OpenMetadata,再启用其 MCP 应用,并把 MCP 客户端连到 `/mcp`。 - 把 MCP 当作“可治理的入口”:血缘、负责人、术语表、标签/分类、质量结果等都能变成 agent 可调用的上下文。 - 当用户不知道具体资产名时,可结合其 Semantic Search 做自然语言发现与定位。 ### Source-backed notes - README 明确写到 OpenMetadata 包含 MCP server,并提供 MCP 文档链接。 - README 的 Quickstart 区域给出 MCP 端点格式:`https:///mcp`。 - README 列出 MCP tools/workflows,并提到可通过 MCP 做 Semantic Search。 ### FAQ - **不用 MCP 也能用吗?**:能。MCP 是其中一种集成方式;平台本身也有 API 与 UI。 - **MCP 带来什么价值?**:README 把 MCP 定位为:让助手/agent 通过自然语言与元数据图谱交互。 - **第一步应该先做什么?**:先把采集与基础治理(负责人、术语表、质量)做起来,这样 MCP 的回答才有可靠上下文。 ## Source & Thanks > Source: https://github.com/open-metadata/OpenMetadata > License: Apache-2.0 > GitHub stars: 13901 · forks: 2095 --- Source: https://tokrepo.com/en/workflows/openmetadata-governed-data-context-via-mcp Author: MCP Hub