MCP Peek Pricing 2026

MCP Peek is a freemium ai agent infrastructure tool that offers 6 key features including stdio and HTTP/SSE transports, OAuth and bearer auth, Live per-tool token counts. MCP Peek is a desktop application for inspecting and debugging Model Context Protocol servers directly, with no model sitting in the middle. Asking an assistant to try a tool shows you what the model did with it — paraphrased and one step removed; MCP Peek connects straight to the server so you read the actual request, the actual response, and the actual timing. It speaks both stdio and HTTP/SSE transports and handles OAuth or bearer auth, then browses everything a server exposes: tools, resources, resource templates and prompts, each rendered with a live token count so you can see what a server costs your context window before you wire it into an agent. Tool inputs can be filled from a generated form or as raw JSON, and results render as table, JSON, Markdown, HTML, image, text or raw. The Pro tier adds a proxy mode: point Claude or Cursor at MCP Peek instead of the server, and every call your agent makes flows through a live capture log with topology view, per-flow detail, status codes, latency and token counts. The developer is explicit that this is a local debugging passthrough and not a production gateway, hub or multiplexer. It runs on macOS, Windows and Linux.

✅ MCP Peek is a good choice if you need:

• •stdio and HTTP/SSE transports
• •OAuth and bearer auth
• •Live per-tool token counts
• •Form or raw-JSON tool invocation
• •Proxy mode with live flow log (Pro)

💡 Value Assessment

With a free tier available, MCP Peek is an easy recommendation for anyone looking to try ai agent infrastructure tools without financial commitment. The paid plans offer good value for power users who need the additional features and higher usage limits.