MCP Inspector

MCP Inspector
A tool worth having when trying out and debugging servers.
A UI tool for trying out and debugging MCP servers before attaching them to an AI assistant.
What is MCP Inspector and who is it for?
The OPCAIOS tools library files MCP Inspector under MCP: it is an MCP server. Short label: GitHub. A UI tool for trying out and debugging MCP servers before attaching them to an AI assistant. MCP Inspector is a developer tool for testing and debugging MCP servers. You connect Inspector to a server, view its list of tools, resources and prompts, and try calling each tool to make sure the server works correctly before handing it to users. Best suited to: Developers building and testing MCP servers. According to OPCAIOS, the editorial score of MCP Inspector is 4.0 out of 5: A tool worth having when trying out and debugging servers.
- Platform: Node.js (run via npx), web, terminal.
- Difficulty to get started: intermediate.
Is MCP Inspector free, and what does it cost?
MCP Inspector is free to use. Free, open source (moving from MIT to Apache-2.0). Additionally, prices are set by the provider and can change, so compare with the official MCP Inspector site before you pay. To compare costs, the MCP group of the library also includes Model Context Protocol (free), MCP Servers (free), and MCP specification (free).
What are the main features of MCP Inspector?
According to OPCAIOS, the standout strengths of MCP Inspector are visual tool calling, multiple modes, and fits automation, among others. Additionally, the list below explains each of the 5 recorded features.
- Visual tool calling: The web interface lets you view and try the server's tools, resources and prompts.
- Multiple modes: Web, CLI, TUI and experimental mcpdo in the same package.
- Fits automation: CLI mode works in scripts, CI and agent loops.
- Runs without installation: Launch directly via npx.
- Secrets stored in the keychain: OAuth tokens and secrets are stored in the operating system keychain by default when available.
When should you use MCP Inspector?
MCP Inspector is used for 4 kinds of work. When using MCP Inspector, the most typical case is checking a server before handover. Additionally, the main audience is developers building and testing MCP servers.
- Checking a server before handover: The IT team tries each tool to confirm the internal server returns correct results.
- Debugging connections: Determine whether the fault is in the server or the AI app when the assistant cannot call a tool.
- Automated testing: Add CLI mode to the CI pipeline to catch bugs each time the server is updated.
- Evaluating third-party servers: See which tools a public server exposes before allowing its use in the company.
How do you get started with MCP Inspector?
When using MCP Inspector for the first time, you go through 5 steps, which OPCAIOS rates as intermediate difficulty. The first step is to open the web interface, and the last is to call and read results. The steps are listed in order below.
- Open the web interface. Run Inspector in default mode to connect to and try a server.
- Use CLI mode. Use it when you need to run checks by script or in CI.
- Use TUI mode. An interface right in the terminal for people comfortable with the command line.
- Configure the server to connect. Declare the server following the MCP server configuration guide in the repo's documentation.
- Call and read results. Select each tool, enter sample parameters and check the response before real use.
What is MCP Inspector?
MCP Inspector is a developer tool for testing and debugging MCP servers. You connect Inspector to a server, view its list of tools, resources and prompts, and try calling each tool to make sure the server works correctly before handing it to users.
Everything is in a single @modelcontextprotocol/inspector package with several modes: a web interface (default), a CLI for scripting and running in CI, a TUI terminal interface, and the experimental mcpdo for running multiple commands over a named connection.
Inspector manages OAuth tokens, client secrets and the server's environment variables; by default it stores them in the operating system keychain if available, and on machines without a keychain it stores them in an unencrypted file unless you provide a key.
How much does MCP Inspector cost?
Free, open source (moving from MIT to Apache-2.0)
Prices are set by the provider and can change; check the official site for the latest.
Key features of MCP Inspector
Visual tool calling
The web interface lets you view and try the server's tools, resources and prompts.
Multiple modes
Web, CLI, TUI and experimental mcpdo in the same package.
Fits automation
CLI mode works in scripts, CI and agent loops.
Runs without installation
Launch directly via npx.
Secrets stored in the keychain
OAuth tokens and secrets are stored in the operating system keychain by default when available.
What is MCP Inspector used for?
Checking a server before handover
The IT team tries each tool to confirm the internal server returns correct results.
Debugging connections
Determine whether the fault is in the server or the AI app when the assistant cannot call a tool.
Automated testing
Add CLI mode to the CI pipeline to catch bugs each time the server is updated.
Evaluating third-party servers
See which tools a public server exposes before allowing its use in the company.
How to use MCP Inspector
- 1
Open the web interface
Run Inspector in default mode to connect to and try a server.
$npx @modelcontextprotocol/inspector - 2
Use CLI mode
Use it when you need to run checks by script or in CI.
$npx @modelcontextprotocol/inspector --cli - 3
Use TUI mode
An interface right in the terminal for people comfortable with the command line.
$npx @modelcontextprotocol/inspector --tui - 4
Configure the server to connect
Declare the server following the MCP server configuration guide in the repo's documentation.
- 5
Call and read results
Select each tool, enter sample parameters and check the response before real use.
MCP Inspector video guides
10 - MCP Inspector: Test and Debug your MCP Server Locally
AI Anytime
Build & Test a MCP Server with TypeScript and MCP Inspector
Roy Derks
How to test MCP server using MCP Inspector
Jaydeep Karale
Real-world example: MCP Inspector
A logistics company in Binh Duong has just written an MCP server for looking up shipment status. Before letting customer care staff use it, a developer opens Inspector, calls the shipment lookup tool with a few sample codes and checks the error message when a wrong code is entered. They then add the Inspector CLI command to the CI process to re-check every time the server is changed.
Tips for using MCP Inspector
Check where secrets are stored
On servers without a keychain, secrets sit in an unencrypted file; provide a key or restrict access.
Use test data
Call tools with sample data to avoid accidentally changing real data.
Automate with the CLI
Put CLI mode in CI so you do not have to test manually after every update.
Need a recent enough Node for development
To build from source, the README requires Node 22.19.0 or higher.
MCP Inspector: frequently asked questions
Does Inspector replace an AI app?
No, it is only a testing tool for calling servers, with no conversational AI model.
Do I need to install it first?
You can run it directly with npx.
What is mcpdo?
An experimental CLI that lets you run multiple commands over a named connection through a local background process.
What is the license?
The project is moving from MIT to Apache-2.0; new contributions are under Apache-2.0, see the LICENSE file for details.
Useful links
Want to apply this in your business?
OPCAIOS looks at your workflows, picks the right tools and sets them up with your team. The first call is free.
