Pytest Mcp Server

What is Pytest Mcp Server?

Pytest Mcp Server is an MCP server for debugging pytest failures using systematic debugging principles. It provides tools to register, analyze, and resolve test failures, and is production‑ready for use with MCP clients.

How to use Pytest Mcp Server?

Install globally via npm install -g pytest-mcp-server. Add to an MCP client (e.g., Claude Desktop) with command pytest-mcp-server start or run as HTTP server using pytest-mcp-server start-http. Data directory is configurable via the DATA_DIR environment variable (default ./data).

Key features of Pytest Mcp Server

• Register and list test failures
• Get detailed failure information
• Apply systematic debugging principles
• Analyze failure patterns
• Generate debugging prompts for LLMs
• Get pytest documentation guidance

Use cases of Pytest Mcp Server

• Debug a failing test by registering its details and applying debugging principles.
• Analyze multiple test failures to identify recurring patterns.
• Automate debugging prompts for LLMs based on test failure data.
• Integrate with Claude Desktop or other MCP clients for interactive debugging.

FAQ from Pytest Mcp Server

What runtime does Pytest Mcp Server require?

Node.js 18.19.0 or higher and npm or yarn.

Where does the server store data?

Data is stored in JSON files (failures.json and debug_sessions.json) in the DATA_DIR directory (default ./data).

Is the server production-ready?

Yes, it is fully functional and has been tested with MCP Inspector and real MCP clients. All 8 tools work correctly.

What known issues exist?

mcp-framework 0.2.13 returns validation errors with "type": "error" instead of the MCP‑compliant "type": "text". This only affects the MCP Testing Framework – real usage is unaffected.

How can I configure it for Claude Desktop?

Add a JSON entry to your Claude Desktop MCP configuration with command pytest-mcp-server and arguments ["start"]. You can set DATA_DIR in the environment variables.