MCP Inspector

Visit Website

Introduction

MCP Inspector is the developer tool maintained in the Model Context Protocol organisation for inspecting MCP servers. It ships as a single package, @modelcontextprotocol/inspector, with four ways to inspect a server.

The web client is a Vite, React and Mantine single-page app with a Node backend, the default when you run npx @modelcontextprotocol/inspector. The CLI is a scriptable command-line client for automation, CI and fast agent feedback loops, started with --cli. The TUI is an interactive terminal UI built with Ink, started with --tui. The experimental mcpdo is a connection CLI: you connect once, then run many commands against the named connection through an implicit local daemon. The first three run through one global mcp-inspector binary, and mcpdo is a second binary in the same package.

Inspector lets you connect to a server, list and call tools, read resources and prompts, and watch the protocol traffic, which makes it the first stop when an MCP server misbehaves in a client.

Security handling deserves attention. The Inspector manages secrets, including OAuth tokens, OAuth client secrets and stdio env: values, and stores them in the OS keychain by default where available. On machines without a keychain, such as Linux without libsecret, headless or SSH sessions, Termux or containers, they are saved to ~/.mcp-inspector/secrets.json, unencrypted unless you supply a key; the docs explain how to restore a keychain, encrypt the file or keep secrets in memory only.

The repository is now on the v2 line. Active development happens on v2/main, merged into main at milestone releases and published to npm’s latest tag; legacy v1 receives security fixes only and is available as @modelcontextprotocol/inspector@v1-latest. A migration guide covers CLI flag changes, the new --config versus --catalog split and the Node engine bump; v2 requires a recent patched Node release on each supported line (22, 24 or 26).

MCP servers are software like any other, but they have an unusual consumer: a model that reads tool names and descriptions and decides what to call. That makes two kinds of testing necessary: protocol-level checks that requests, responses, auth and errors behave correctly, and behavioural checks that real models actually choose and use the tools as intended. Inspector handles the protocol side; for behavioural checks across models and clients, combine it with an evaluation tool.

The source code is at modelcontextprotocol/inspector on GitHub under the MIT, transitioning to Apache-2.0 licence, with about 11,300 stars; the most recent commit was on 4 November 2026.

In CI, the CLI mode can list tools and call a few with known inputs after each build, failing the pipeline if a tool disappears, changes its schema unexpectedly or returns an error. That simple smoke test catches many regressions before an MCP client ever sees them.

For manual debugging, the web UI is the fastest way to see exactly what a server returns, including error payloads that clients sometimes hide from users, which makes it easier to tell a server bug from a client bug.

Remember to clear stored secrets on shared machines and CI runners, especially where the plain-file fallback is used, so tokens from debugging sessions do not linger.

Features

1. Web UI: Interactive inspection.
2. CLI: Scriptable for CI and agents.
3. TUI: Ink-based terminal UI.
4. mcpdo: Experimental persistent connection client.
5. Tools, resources and prompts: Call and read everything a server exposes.
6. OAuth support: Manages tokens and client secrets.
7. Keychain storage: Secrets in the OS keychain.
8. v1 compatibility: Legacy line still available.

Pros

1. Official: Maintained in the MCP organisation.
2. Zero install: Runs with npx.
3. Free to use: Open source under the MIT, transitioning to Apache-2.0 licence, so the core tool costs nothing to use.
4. Actively maintained: The repository had commits as recent as 4 November 2026.
5. Community: About 11,300 GitHub stars.

Cons

1. Secret storage caveat: Plain file fallback without a keychain.
2. Node version requirements: v2 needs recent Node releases.
3. v2 migration: CLI flags changed.

Usage and Links (Example)

npx @modelcontextprotocol/inspector          # web UI (default)
npx @modelcontextprotocol/inspector --cli    # CLI for scripts and CI
npx @modelcontextprotocol/inspector --tui    # terminal UI
npx -p @modelcontextprotocol/inspector mcpdo --help
npx @modelcontextprotocol/inspector@v1-latest   # legacy v1

Resources:

• MCP Inspector website

Pricing

MCP Inspector is free and open source under the MIT, transitioning to Apache-2.0 licence.

Recommended for?

1. Anyone building an MCP server.
2. CI pipelines smoke-testing MCP servers with the CLI.
3. Agent developers debugging tool calls.

Conclusion

MCP Inspector is the essential first tool for MCP server development.

Use the CLI in CI and pair it with behavioural evals.

On CheckOps, MCP Inspector is filed under the API & contract testing, Integration testing and Release & CI gates stages. Related tools listed here include Playwright MCP, Chrome DevTools MCP, Postman, Promptfoo, MCP Selenium and HTTPie. When comparing them, weigh the licence, the languages and platforms each supports, whether it is self-hosted or managed, and how cleanly it slots into your existing CI pipeline. Details on this page were checked against the official website, documentation and repository on 11 November 2026.