Open-source testing and evaluation platform for MCP servers, for inspecting, chatting with, and debugging MCP servers, MCP apps, and ChatGPT apps.
standardbeagle/mcp-tui
Terminal UI and CLI for testing, debugging and automating Model Context Protocol (MCP) servers over stdio, SSE and HTTP.
Open-source alternative to:Postman
- Stars
- 12
- Forks
- 2
- Open issues
- 15
- Last commit
- 4 days ago
What is it?
mcp-tui lets you connect to a Model Context Protocol server from the terminal, pick a tool, fill in a form generated from its input schema, and run the call. Every call can also be copied as an equivalent CLI command, so a manual check can move into a shell script or CI job.
When a connection fails, the tool reports the specific cause, such as an HTTP 404 on the wrong path, a transport mismatch, a stdio server writing logs to stdout, or a missing OAuth sign-in, along with a suggested fix. It also logs every JSON-RPC message in both directions, HTTP timings and the OAuth flow step by step, with credentials masked.
The project includes a `verify` command for spec and security probes and a `conform` command that runs protocol scenarios and writes a JUnit report. It is built on the official MCP Go SDK and is distributed through `go install` and npm.
Key features
- 01Generates a tool-call form from each tool's input schema in the TUI.
- 02Runs the same calls from the CLI with JSON output for jq.
- 03Explains connection failures with causes and fixes.
- 04Logs every JSON-RPC message, HTTP timing and OAuth step.
- 05Exports TUI sessions as replayable shell scripts.
- 06Writes JUnit conformance reports for CI.
Use cases
- Manually testing a new MCP server's tools before wiring it into an AI client.
- Debugging why an MCP server fails to connect over HTTP or stdio.
- Adding MCP server conformance checks to a CI pipeline.
- Scripting repeatable tool calls against an MCP server from the shell.
Quick start
- 1
Install with Go
go install github.com/standardbeagle/mcp-tui@latest - 2
Or install with npm
npm install -g @standardbeagle/mcp-tui - 3
Open the TUI against an HTTP MCP server
mcp-tui --url http://localhost:8080/mcp
Things to know
- Early-stage project with only 12 stars, so adoption and track record are limited.
Topics
FAQ
- What is mcp-tui?
- Terminal UI and CLI for testing, debugging and automating Model Context Protocol (MCP) servers over stdio, SSE and HTTP. mcp-tui lets you connect to a Model Context Protocol server from the terminal, pick a tool, fill in a form generated from its input schema, and run the call. Every call can also be copied as an equivalent CLI command, so a manual check can move into a shell script or CI job.
- Is mcp-tui free to use? What license does it have?
- mcp-tui is open source under the MIT license. Check the license terms before using it in commercial products.
- Is mcp-tui still maintained?
- Actively maintained. The last commit to mcp-tui was on Oct 18, 2026.
- What can I use mcp-tui for?
- Manually testing a new MCP server's tools before wiring it into an AI client. Debugging why an MCP server fails to connect over HTTP or stdio. Adding MCP server conformance checks to a CI pipeline. Scripting repeatable tool calls against an MCP server from the shell.
- How do I install mcp-tui?
- Install with Go: go install github.com/standardbeagle/mcp-tui@latest Or install with npm: npm install -g @standardbeagle/mcp-tui Open the TUI against an HTTP MCP server: mcp-tui --url http://localhost:8080/mcp
- Where can I get mcp-tui?
- The source code and installation instructions are on GitHub: https://github.com/standardbeagle/mcp-tui
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A visual and command-line developer tool for testing and debugging Model Context Protocol (MCP) servers, with web, CLI, and terminal UI modes.
Hurl is a command line tool that runs and tests HTTP requests written in a plain text file format.
A terminal UI for viewing, checking, installing and updating agent skills across Claude Code, OpenCode, Codex, Cursor and other coding agents.
A CLI and web tool that packs an entire code repository into one AI-friendly file for feeding to LLMs like Claude or ChatGPT.
An MCP server that connects AI agents to Exa for web search, page content fetching, and multi-step research.