Twisted66/ai-testing-mcp

AI Testing MCP Server

An AI-powered testing automation MCP server for Claude Code that provides TestSprite-like functionality.

Features

Core Testing Tools

• Code Analysis: Analyze project structure and identify testable components
• Test Generation: AI-powered unit, integration, and E2E test generation
• Test Execution: Run tests with framework auto-detection
• Results Analysis: Intelligent test result analysis with fix suggestions
• Framework Setup: Initialize testing frameworks in projects

Supported Languages & Frameworks

• JavaScript/TypeScript: Jest, Mocha, Vitest
• Python: Pytest
• Go: Built-in testing
• Frameworks: React, Vue, Express, FastAPI, Flask

Installation

1. Clone or create the project:

git clone <repo> ai-testing-mcp
cd ai-testing-mcp

2. Install dependencies:

npm install

3. Build the project:

npm run build

Configuration

For Claude Code

Add to your Claude Code MCP configuration:

{
  "mcpServers": {
    "ai-testing": {
      "command": "node",
      "args": ["dist/index.js"],
      "cwd": "/absolute/path/to/ai-testing-mcp"
    }
  }
}

Usage

Once configured, you can use these tools in Claude Code:

1. Analyze Codebase

Use the analyze_codebase tool on my project at /path/to/project

2. Generate Unit Tests

Generate unit tests for src/utils.js using Jest framework

3. Run Tests

Run all tests in my project with coverage

4. Analyze Test Results

Analyze these test failures and suggest fixes: [paste test output]

Available Tools

analyze_codebase

• Purpose: Scan project structure, detect languages/frameworks, find API endpoints
• Input: projectPath - Path to project directory
• Output: JSON with language, framework, API endpoints, testable components

generate_unit_tests

• Purpose: Create unit tests for functions/components
• Input:
  • filePath - File to test
  • functionName - Specific function (optional)
  • testFramework - Framework to use
• Output: Generated test code

generate_integration_tests

• Purpose: Create API endpoint tests
• Input:
  • projectPath - Project directory
  • apiSpec - API documentation (optional)
• Output: Integration test code

run_tests

• Purpose: Execute tests with auto-detection
• Input:
  • projectPath - Project directory
  • testPattern - Specific tests to run (optional)
  • framework - Force specific framework (optional)
• Output: Test results with coverage

analyze_test_results

• Purpose: Analyze test output and provide insights
• Input: testOutput - Raw test results
• Output: Analysis with suggestions

suggest_fixes

• Purpose: Generate fix suggestions for failing tests
• Input:
  • failureOutput - Test failure output
  • sourceCode - Source code (optional)
  • testCode - Test code (optional)
• Output: Suggested fixes

setup_testing_framework

• Purpose: Initialize testing in a project
• Input:
  • projectPath - Project directory
  • framework - Testing framework
  • language - Programming language
• Output: Setup instructions/files

Example Workflow

1. Analyze your project:

   Use analyze_codebase on ./my-app
   

2. Generate tests for a component:

   Generate unit tests for src/components/Button.jsx using Jest
   

3. Run the generated tests:

   Run tests in ./my-app matching *Button*
   

4. Fix any failures:

   Analyze these test failures and suggest fixes: [paste output]
   

Benefits Over Manual Testing

• Speed: Generate comprehensive test suites in seconds
• Coverage: Automatically identifies all testable components
• Intelligence: AI-powered test case generation with edge cases
• Integration: Works with existing test frameworks and CI/CD
• Analysis: Intelligent failure analysis with fix suggestions

Development

Build

npm run build

Development Mode

npm run dev

Adding New Features

The server is modular - add new tools by:

1. Adding tool definition to ListToolsRequestSchema handler
2. Adding tool implementation to CallToolRequestSchema handler
3. Creating helper methods as needed

Comparison to TestSprite

License

MIT