← Back to directory
M

mcp-use Inspector

Community
Framework / SDK
An inspector for testing and debugging MCP servers and apps.

This is a framework for building MCP servers or agents, not a server you connect a client to.

Category
Dev Tools #158 of 438
Stars
★ 10.8k Very popular
Runtime
Node.js
Credentials
No credential needed
License
MIT
Last commit
57FMRS · C

This is a development and debugging component in the mcp-use repository for testing and previewing MCP servers and MCP Apps. The source provides no standalone tool list, transport declaration, client compatibility list, or installation configuration, so those fields remain empty.

Strongest · Documentation 14/20 Weakest · Reliability 8/20

Reliability
8/20
Security and permissions
9/20
Maintenance
13/20
Documentation
14/20
Setup experience
13/20
Why each score
Reliability 8/20
The manifest shows a TypeScript Inspector compilation configuration including server, CLI, and LLM sources, but it cannot verify startup, the MCP initialization and tool-list handshake, or whether declared tools match actual behavior. The README provides plausible framework examples, but there is no committed test, CI coverage, or error-handling evidence. This warrants a conservative deduction and keeps the score below the static-review cap of 12.
Security and permissions 9/20
The supplied material shows no malware, credential theft, covert exfiltration, or other security red line, so this is not a 0-4 score. However, the framework covers MCP servers, clients, agents, Inspector, deployment, and external services without sufficient evidence of authentication, credential storage, least privilege, data boundaries, or confirmation for dangerous actions. Examples also include file management, dataset deletion, and deployment capabilities whose controls are not established.
Maintenance 13/20
The repository is not archived, uses the MIT license, and exposes SECURITY.md, contribution guidance, GitHub Issues, documentation, and community support channels, with named core contributors. These are positive maintenance signals, but the prompt provides no verifiable release history, commit cadence, dependency-update practice, issue-response data, or security-response record. The 119 open issues do not establish maintenance quality, so the score remains well below full marks.
Documentation 14/20
The README covers TypeScript and Python installation, servers, MCP Apps, Inspector, agents, clients, deployment, and multiple examples, with documentation and security links. Points are deducted because authentication, permissions, limits, costs, compatibility, troubleshooting, and the concrete behavior of the target Inspector server path are not sufficiently documented. The README also contains the apparent “MPC App” typo.
Setup experience 13/20
The material provides an npx scaffold, pip installation, TypeScript and Python quickstarts, an automatic Inspector URL, client configuration examples, and a standalone Inspector command, making the basic path fairly clear. Points are deducted because runtime and dependency prerequisites, production authentication, mainstream-client connection details, platform differences, and the actual build flow for the target manifest are not source-verifiable. Static review also lacks test or CI evidence.

Static review · not runListed 2026-07-29

Read the FMRS scoring method →

Fit and risk

What it can accessUses the network

Best for

  • TypeScript or Python developers building MCP servers
  • Teams validating MCP App behavior
  • Users verifying server connections during development

Not for

  • End users seeking business-data tools
  • Production scenarios requiring documented authentication, authorization, or operational guarantees
  • Users looking for ready-made business tools rather than development tooling

Required permissions

  • The source does not specify required system permissions, account permissions, or data-access scope.
  • When connecting to an external server, actual permissions depend on the inspected MCP server.

Risks and side effects

  • The source does not specify an authentication or authorization model for the Inspector.
  • Inspecting a server may expose the tools, resources, or debugging information provided by that server.
  • Connecting the Inspector to an untrusted server may introduce unknown risks.

Setup

Before you start

Runtime:Node.js

The Inspector is auto-included when using mcp-use server.listen() or server.run(). It can also be launched with npx @mcp-use/inspector --url http://localhost:3000/mcp to inspect a specified server, or accessed through the online Inspector.

Check that it works

Run npx -y create-mcp-use-app@latest, then npm run dev in the generated project, and open http://localhost:3000/mcp/inspector; the get-weather tool should be listed and callable there.

Troubleshooting

  1. Confirm that the inspected server is running.
  2. Verify that the MCP endpoint is correct, such as http://localhost:3000/mcp.
  3. After using server.listen() or server.run(), check the corresponding /inspector address.
  4. Review server logs and Inspector-reported connection or tool errors.

Things to try

Once connected, you can ask your AI assistant things like:

  • Build an MCP server with mcp-use that returns weather for a city
  • Add a React weather-card view bound to the get-weather tool
  • Invoke the get-weather tool with city=Tokyo in the Inspector

Use cases

Interactively test MCP servers
Debug MCP Apps and server connections
Preview capabilities exposed by an MCP server

Overview

mcp-use is a fullstack MCP framework for building MCP servers, MCP Apps, clients, and AI agents. This server path corresponds to the mcp-use Inspector, which provides interactive testing and debugging for MCP servers and can also inspect standalone MCP server endpoints.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →

Source revision eec124caf602 Data synced 2026-10-19 Read the FMRS scoring method