Skip to content
MCPJam Inspector

MCPJam InspectorMCP Server

Released
2 views
v3.3.1
Apache-2.0

Summary

Testing and evaluation workbench for MCP server authors — inspect tools and prompts, debug OAuth step by step, and score behaviour across 16 client configurations before shipping.

Features

  • Playground emulating 16 client configurations and 170+ models with full JSON-RPC traces
  • Step-by-step OAuth and EMA debugger with conformance checks across four protocol revisions
  • Chrome DevTools-style widget emulator for OpenAI Apps SDK and MCP app UIs
  • Eval test cases with expected tool calls, scored for accuracy and latency over time
  • CLI and SDK for running conformance, E2E and OAuth checks in CI/CD
  • Workspaces with real-time team sync so a team tests one shared server config

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it.
  3. Restart the agent — the "MCPJam Inspector" tools will be available.
{
  "mcpServers": {
    "mcpjam-inspector": {
      "command": "npx",
      "args": [
        "@mcpjam/inspector@latest"
      ]
    }
  }
}

Description

ChatGPT, Claude and Cursor each read the same MCP server differently, and a server that behaves in one client can quietly break in another. MCPJam Inspector exists to find that gap before users do.

Inspecting and debugging

The Playground is a cross-client chat surface that emulates real client UIs, tool calls and skills, with full traces on every JSON-RPC message. It renders OpenAI Apps SDK and MCP app UIs and includes a Chrome DevTools-style widget emulator — switch between Desktop, Tablet and Mobile, test locale changes, CSP permissions, light and dark mode, hover versus touch, and safe-area insets without deploying anywhere. Server Debugging covers the manual path: run tools, resources, templates, prompts and elicitation flows directly and read the raw JSON-RPC log.

The OAuth debugger

Authorization is where MCP integrations most often fail, and MCPJam visualises OAuth and EMA exchanges step by step rather than reporting one opaque error. It runs guided conformance checks across protocol revisions 03-26, 06-18, 11-25 and 2026-07-28, covering dynamic client registration, client pre-registration and CIMD.

Evaluations and regression gating

Test cases declare the tool calls a prompt should produce; MCPJam runs them across models and tracks accuracy, latency and tool-call performance over time. The CLI and SDK make the same checks runnable in GitHub Actions or any pipeline, so a conformance or behaviour regression fails the pull request instead of reaching production. Workspaces share server groups across a team with real-time sync, so everyone tests against one configuration.

Running it

There is a hosted app at app.mcpjam.com that needs no install, and npx @mcpjam/inspector@latest runs it locally — necessary for local stdio servers. Apache-2.0 licensed and open source.

Related MCP Servers

New

Arm's official MCP server for porting and tuning software on Arm — semantic docs search, codebase migration scanning, container architecture checks and LLVM-MCA assembly analysis.

New

Official AntV MCP server that turns a description of your data into a rendered chart — 26 visualization types from bar and line to sankey, mind-map, fishbone and geographic maps.

MCP: Headroom

by Headroom Labs

New

Context-compression MCP server that shrinks tool outputs, logs, files and RAG chunks before they reach the model, cutting 60-95% of tokens on JSON payloads.

MCP: Engram

by Gentleman Programming

New

Persistent memory for coding agents as a single Go binary — SQLite and FTS5 behind MCP, an HTTP API, a CLI and a TUI, with no Node, Python or Docker required.

Browse all MCP servers →