MCPJam/inspector

Testing and evaluation platform to chat, inspect, and debug MCP servers, MCP apps, and ChatGPT apps.

What it solves

MCPJam is a testing and evaluations platform designed for developers building Model Context Protocol (MCP) servers. It solves the problem of inconsistent behavior across different AI clients (like Claude, ChatGPT, and Cursor) by providing a unified environment to inspect, debug, and validate how these clients interact with a server's tools, prompts, and resources.

How it works

The platform provides a suite of tools for both manual and automated testing:

  • Playground & Chat: A cross-client chat interface that emulates various UI configurations and allows developers to compare how different frontier models handle tool calls and server interactions.
  • Server Debugging: Provides full JSON-RPC observability, allowing developers to manually run tools, resources, and prompts to ensure they return the correct data.
  • OAuth Debugger: A guided tool to ensure MCP OAuth conformance across multiple protocol versions.
  • Evals & CI/CD: Allows the creation of test cases with expected tool calls that can be run programmatically via a CLI or SDK to catch regressions in CI/CD pipelines.

Who it’s for

Developers who are building and shipping MCP servers and need to ensure their integrations work reliably across a diverse range of AI agents and LLM clients.

Highlights

  • Cross-Client Emulation: Test against 16 client configurations and 170+ models.
  • Full Observability: Detailed traces of every JSON-RPC message and OAuth exchange.
  • Automated Evals: Define expected tool calls and track accuracy metrics over time.
  • Flexible Deployment: Available as a hosted web app, desktop application, or via npx for local development.
  • Collaborative Workspaces: Shared server groups with real-time sync for team testing.

Related

  • Project
    modelcontextprotocol/inspector

    A developer tool for inspecting and debugging Model Context Protocol (MCP) servers via Web, CLI, and TUI interfaces.

  • Project
    microsoft/DebugMCP

    An MCP server that grants AI coding agents autonomous control over the VS Code debugger to set breakpoints, step through code, and inspect variables locally.

  • Project
    IBM/mcp

    A collection of Model Context Protocol (MCP) servers and tools that enable AI agents to securely interact with IBM products and enterprise resources.

  • Project
    pathintegral-institute/mcpm.sh

    A CLI tool for managing Model Context Protocol (MCP) servers, providing global configuration, server discovery, and easy integration across multiple AI clients.

  • Project
    googleapis/mcp-toolbox

    An open-source MCP server that connects AI agents and IDEs to enterprise databases via prebuilt generic tools or a custom framework for secure, structured data access.