MCP.so
Sign In

mcp-server-openmetadata

@tpavelek

About mcp-server-openmetadata

No overview available yet

Basic information

Category

Other

License

MIT

Runtime

python

Transports

stdio

Publisher

tpavelek

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "om-glossary-mcp-server": {
      "command": "python",
      "args": [
        "src/server.py"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is mcp-server-openmetadata?

A Model Context Protocol (MCP) server that wraps OpenMetadata's REST API, enabling standardized interaction with OpenMetadata through MCP clients.

How to use mcp-server-openmetadata?

Configure via environment variables (token or basic auth) and add to Claude Desktop's claude_desktop_config.json using uvx or uv. Run manually with python src/server.py with optional --port and --transport flags.

Key features of mcp-server-openmetadata

  • Implements MCP protocol for OpenMetadata integration
  • Supports token-based and basic authentication
  • Provides CRUD operations for tables (list, get, create, update, delete)
  • Runs via stdio (default) or SSE transport
  • Configurable server port (default 8000)

Use cases of mcp-server-openmetadata

  • Manage OpenMetadata tables through MCP clients like Claude Desktop
  • Automate metadata operations using standardized MCP tools
  • Integrate OpenMetadata into larger MCP-based workflows

FAQ from mcp-server-openmetadata

What is MCP?

MCP stands for Model Context Protocol, a standardized protocol for interacting with external systems. This server wraps OpenMetadata's REST API to expose it via MCP.

How do I authenticate to OpenMetadata?

Set OPENMETADATA_HOST and either OPENMETADATA_JWT_TOKEN (recommended) or OPENMETADATA_USERNAME and OPENMETADATA_PASSWORD as environment variables.

What features are currently implemented?

Only table operations are implemented: list, get (by ID and name), create, update, and delete. All other endpoints (databases, services, teams, users, search, tags, lineage, glossaries, usage) are not yet available.

How do I run the server manually?

Run python src/server.py with optional --port (default 8000) and --transport (stdio or sse, default stdio).

What transports are supported?

The server supports both stdio (default) and SSE transport types.

Comments

More Other MCP servers