​
What is MCPcat?

MCPcat provides a comprehensive understanding of how users (and agents) engage with and interact with your MCP server implementations. By combining session replay with product analytics and technical usability issues, MCPcat offers actionable insights into what you can do to deliver the best experience for every AI user. MCPcat seamlessly integrates with your existing observability platforms without requiring an account.

​
What can MCPcat help with?

Debug Tool Calls

Trace and debug every tool call made to your MCP server. Identify errors, performance issues, and unexpected behavior in real-time.

Session Replay

Record and replay user sessions to debug issues and understand usage patterns. Track behavior across multiple sessions.

One-Line Setup

Add analytics with a single line of code. SDKs available for TypeScript and Python.

Works With Your Stack

Connect seamlessly with OpenTelemetry, Datadog, Sentry, and other observability platforms you already use.
MCPcat’s number one goal is to help product teams building MCP servers answer the following questions:
  1. Why are people (or agents) using your MCP servers?
  2. Who are the people (or agents) using your MCP servers?
  3. How are the people (or agents) using your MCP server getting stuck?