Track MCP, written as TrackMCP, is analytics and observability for Model Context Protocol servers. It shows which clients connect, which tools agents use, where calls fail, and whether workflows complete.

How do I start using Track MCP?

Install the TrackMCP TypeScript or Python SDK, wrap the MCP server at its boundary, create a workspace key, and make one real tool call. The first event then appears in the TrackMCP dashboard.

Track MCP

Track MCP servers from connection to outcome.

Track MCP, written as TrackMCP, is the analytics and observability layer for teams shipping Model Context Protocol servers. See who connects, what agents call, where a workflow breaks, and what to improve next.

Start measuring Read the docs

See who connects

Identify the MCP clients, environments, and sessions creating real demand.

Understand tool use

See which tools agents discover, choose, retry, and leave unused.

Improve the workflow

Find slow calls, silent application errors, and the point where work stopped.

How Track MCP works

Add the SDK at the server boundary. TrackMCP captures protocol methods, tool calls, sessions, clients, errors, latency, retries, and workflow signals while leaving your tools and responses unchanged.

Install the TypeScript or Python SDK
Wrap your existing MCP server
Make one real tool call and inspect the result

MCP server analytics Open the quickstart

Frequently asked questions

What is Track MCP?

Track MCP is the spaced way to write TrackMCP, the analytics and observability product for Model Context Protocol servers.

Who is Track MCP for?

Track MCP is for teams that build or operate MCP servers and need to understand adoption, reliability, tool usage, and workflow outcomes.