by MCPJam
MCPJam is an open-source testing and evals platform for MCP server developers. It lets you inspect tools, debug OAuth, and check how your server behaves across many AI clients.
Web appContact for pricing
Step 1: Run npx @mcpjam/inspector@latest, or open the hosted web app.
Step 2: Connect your MCP server using a config shaped like an mcp.json file.
Step 3: Call tools and prompts in the Inspector and read the trace.
Step 4: Use the OAuth debugger if authorization fails.
Step 5: Create eval test cases and run them across clients and models.
Step 6: Add the eval suite to GitHub Actions to gate pull requests.
The core is open source and free, and runs locally or in your own CI with no signup. Paid plans add shared credits, higher eval limits, SSO and support.
Yes. If an MCP client can connect to your server, MCPJam can test it, regardless of SDK, framework or deployment.
No. It is pre-production only and exercises your server the way a client would.
It uses evaluation rather than brittle assertions, scoring tool choice, arguments and goal completion across repeated runs.
BoltBolt is an AI builder for websites, web apps, mobile apps and prototypes that runs in the browser. You describe what you want in a prompt and publish the result to a live address.
DelysiumDelysium is building a blockchain-based network where AI agents collaborate. Its first agent, Lucy, is an AI-powered Web3 operating system on the YKILY Network.
DifyDify is a platform for building agentic workflows, agents and RAG pipelines on one visual canvas. It runs as a hosted cloud, a private enterprise deployment, or open-source self-hosted software.