MCPJam Inspector

MCPJam Inspector is an open-source testing, inspection, and evaluation platform for developers building Model Context Protocol servers. It connects to servers running over STDIO, SSE, and Streamable HTTP protocols, allowing engineers to probe capabilities, test tool executions, and inspect JSON-RPC messages. Developers and QA engineers use it to diagnose communication failures, validate resource templates, and debug OAuth authorization flows across different specifications. The platform enables multi-server testing, cross-client behavioral evaluations, and side-by-side LLM comparisons against popular models to observe variance across environments like ChatGPT, Claude, and Cursor. With dedicated views for chat sessions, raw message logs, and execution traces, it provides full visibility into agent steps and payload exchanges. MCPJam Inspector can be operated via an interactive web interface, a standalone desktop application, a command-line interface, or an SDK integrated into continuous integration pipelines to prevent regressions before shipping server updates.

Category: AI & LLM Tooling

Tags: debugging, http, sse, stdio, testing

Visit MCPJam Inspector

How to install and configure MCPJam Inspector

Follow these steps to run and use MCPJam Inspector: 1. Ensure you have Node.js 20+ installed if running via the command line, or use the hosted web app. 2. Launch the terminal inspector directly using npx: bash npx @mcpjam/inspector@latest 3. Alternatively, build and run using Docker locally: bash git clone https://github.com/MCPJam/inspector.git cd inspector docker build -t mcpjam/mcp-inspector:local -f mcpjam-inspector/Dockerfile . docker run -p 127.0.0.1:6274:6274 mcpjam/mcp-inspector:local 4. Open the printed localhost URL in your browser to start testing STDIO, SSE, or Streamable HTTP servers. 5. For detailed client configurations and CI/CD setup, refer to the repository at https://github.com/MCPJam/inspector.

What you can do with MCPJam Inspector

• Debugging server tools, prompts, and resources manually while inspecting raw JSON-RPC messages and traces. * Validating OAuth and EMA conformance step-by-step across Dynamic Client Registration and Client ID Metadata Documents. * Comparing prompt responses and tool-call reliability across up to three frontier LLMs side by side. * Automating regression testing, conformance checks, and eval test cases inside GitHub Actions or CI/CD pipelines. * Emulating UI widgets and testing client layouts across desktop, tablet, and mobile viewport presets.

Key facts

• Open Source
• https://github.com/MCPJam/inspector
• AI & LLM Tooling, Developer Tools & Code Intelligence
• debugging, http, sse, stdio, testing

Part of MCP Servers

What is MCPJam Inspector?

MCPJam Inspector is a developer tool and evaluation suite designed to test, debug, and inspect Model Context Protocol servers. It provides real-time JSON-RPC logging, OAuth flow debugging, and evaluation capabilities across different language models and client environments.

How do I install MCPJam Inspector?

You can run MCPJam Inspector locally by executing npx @mcpjam/inspector@latest in your terminal with Node.js 20 or higher. You can also build and run its Docker container locally, download the desktop app for macOS and Windows, or use the hosted web version.

What protocols does MCPJam Inspector support?

MCPJam Inspector supports STDIO, Server-Sent Events (SSE), and Streamable HTTP protocols. The terminal, Docker, and desktop applications support both local STDIO and remote HTTP/S connections, while the hosted web app operates over HTTPS.

Which MCP clients work with MCPJam Inspector?

MCPJam Inspector works as an evaluation and diagnostic tool that emulates or tests behavior across sixteen client configurations, including ChatGPT, Claude, Cursor, and GitHub Copilot, ensuring your server functions reliably across diverse host environments.

Is MCPJam Inspector open source?

Yes, MCPJam Inspector is open source. It is licensed under the Apache License 2.0, allowing developers to inspect the source code, build custom local containers, or contribute directly to the project.