API Lab MCP 🧪

Transform Claude into your AI-powered API testing laboratory. Test, debug, and document APIs through natural conversation.

🌟 Why API Lab MCP?

Stop switching between tools. Stop writing repetitive test scripts. Start having conversations with your APIs.

The Problem: Traditional API testing tools require constant context switching, manual script writing, and repetitive workflows that slow down development.

The Solution: API Lab MCP integrates directly with Claude Desktop and Claude Code, allowing you to test APIs through natural language while you code. No more jumping between Postman, terminal, and your IDE.

What makes it different?

🤖 AI-First Design: Built specifically for Claude's conversational interface
💬 Natural Language: Test APIs by describing what you want, not clicking through UIs
🔄 Zero Context Switching: Stay in your development flow
📚 Auto Documentation: Generate API docs from your test conversations
🎯 Smart Assertions: Claude understands your intent and creates appropriate tests

✨ Key Features

🔐 Universal Authentication - Bearer tokens, API keys, OAuth2, session cookies, and CSRF tokens
📊 Intelligent Response Analysis - Automatic validation of status codes, headers, and JSON paths
⚡ Real-time Performance Metrics - Response times, payload sizes, and Core Web Vitals
🎯 Conversational Testing - Test complex scenarios through natural dialogue
📋 Spec-based Test Generation - Auto-generate test suites from OpenAPI/Swagger docs
🔄 Batch Testing - Run multiple APIs in parallel or sequence
🧪 MCP Server Testing - Test other MCP servers' tools and resources
🌍 Environment Management - Switch between dev, staging, and production seamlessly

🚀 Quick Start (30 seconds)

For Claude Code

# Install globally
npm install -g api-lab-mcp

# Add to Claude Code
claude mcp add api-lab-mcp api-lab-mcp

# Start testing!
# Tell Claude: "Test my API at localhost:3000"

For Claude Desktop

# Install via npm
npm install -g api-lab-mcp

# Then add to Claude Desktop config (see Installation section)

That's it! You're ready to test APIs with Claude.

📸 Live Demo with Claude Desktop

Watch API Lab MCP in action with Claude Desktop - testing real APIs through conversation:

Real Results from Claude Desktop

"Test your APIs through natural conversation - in any language you prefer."

💡 Use Cases

For Development Teams

API Development: Test endpoints as you build them without leaving your IDE
Debugging: Quickly diagnose why an API is failing with intelligent analysis
Documentation: Auto-generate API docs from your test conversations
Onboarding: New team members can test APIs without learning complex tools

For QA Engineers

Test Automation: Generate test suites from OpenAPI specs automatically
Regression Testing: Run batch tests across all endpoints
Performance Testing: Monitor response times and identify bottlenecks
Session Testing: Test complex authentication flows with cookies and tokens

For DevOps

Health Checks: Monitor API availability across environments
Load Testing: Test API performance under various conditions
Environment Validation: Ensure APIs work correctly after deployment
Integration Testing: Test API integrations with third-party services

💬 Real-World Examples

Test a REST API

You: "Test the GitHub API to get my user info"
Claude: *Tests GET https://api.github.com/user with your token*

Debug a Failed Request

You: "Why is my login endpoint returning 401?"
Claude: *Analyzes headers, body, and suggests missing auth token*

Session-based Authentication

You: "Test my app with session cookie JSESSIONID=ABC123"
Claude: *Configures session and tests authenticated endpoints*

Performance Testing

You: "Check if all my endpoints respond under 200ms"
Claude: *Runs performance tests and generates report*

Generate Tests from Swagger

You: "Create tests from my OpenAPI spec at /api/docs"
Claude: *Generates comprehensive test suite automatically*

🔧 Installation

Claude Code Installation

# Method 1: Global install (recommended)
npm install -g api-lab-mcp
claude mcp add api-lab-mcp api-lab-mcp

# Method 2: Using npx (no install needed)
claude mcp add api-lab-mcp "npx -y api-lab-mcp"

# Method 3: Local development
git clone https://github.com/atototo/api-lab-mcp.git
cd api-lab-mcp
npm install && npm run build
claude mcp add api-lab-local "node $(pwd)/dist/mcp/server.js"

Claude Desktop Installation

macOS/Linux Setup

Install the package:

npm install -g api-lab-mcp

Find your config file:

# macOS
open ~/Library/Application\ Support/Claude/

# Linux
open ~/.config/claude/

Edit claude_desktop_config.json:

{
  "mcpServers": {
    "api-lab-mcp": {
      "command": "npx",
      "args": ["-y", "api-lab-mcp"]
    }
  }
}

Restart Claude Desktop

Windows Setup

Install the package:

npm install -g api-lab-mcp

Open config folder:

explorer %APPDATA%\Claude

Edit claude_desktop_config.json:

{
  "mcpServers": {
    "api-lab-mcp": {
      "command": "npx.cmd",
      "args": ["-y", "api-lab-mcp"]
    }
  }
}

Restart Claude Desktop

Docker Setup

FROM node:20-alpine
RUN npm install -g api-lab-mcp
EXPOSE 3000
CMD ["api-lab-mcp", "--http-mode"]

Run with:

docker build -t api-lab-mcp .
docker run -p 3000:3000 api-lab-mcp

Verify Installation

After installation, ask Claude:

"Can you test the API at https://api.github.com?"

If Claude can test the API, installation was successful!

🏗️ For Developers

Local Development Workflow

API Lab MCP supercharges your local development:

Real-time API Testing

"Test POST /api/users on localhost:3000 with mock data"

Smart Debugging

"Why is this endpoint slow? Profile the response time"

Automatic Documentation

"Generate README docs from these test results"

Pre-deployment Validation

"Run health checks on all endpoints before I deploy"

Development Setup

# Clone the repository
git clone https://github.com/atototo/api-lab-mcp.git
cd api-lab-mcp

# Install dependencies
npm install

# Run in development mode
npm run dev

# Build for production
npm run build

# Run tests
npm test

Architecture

api-lab-mcp/
├── src/
│   ├── mcp/              # MCP server implementation
│   ├── core/             # Core business logic
│   ├── http/             # HTTP client layer
│   └── types/            # TypeScript definitions
├── tests/                # Test suites
└── docs/                 # Documentation

📊 How It Compares

Feature	API Lab MCP	Postman	Insomnia	Thunder Client
AI-Powered Testing	✅ Native	❌	❌	❌
Natural Language	✅	❌	❌	❌
Claude Integration	✅	❌	❌	❌
Zero Setup	✅	❌	❌	⚠️
Session Management	✅	✅	✅	✅
OpenAPI Support	✅	✅	✅	⚠️
MCP Protocol	✅	❌	❌	❌
Conversational UI	✅	❌	❌	❌
Auto Documentation	✅	⚠️	⚠️	❌
Free & Open Source	✅	⚠️	⚠️	✅

📚 Documentation

Core Tools

`test_http_endpoint`

Test any HTTP endpoint with full customization:

Supports all HTTP methods (GET, POST, PUT, DELETE, PATCH, etc.)
Custom headers and authentication (Bearer, API Key, Basic Auth)
Request body with JSON, form data, or raw content
Returns detailed response with status, headers, body, and timing

`test_with_assertions`

Advanced testing with intelligent response validation:

JSONPath expressions for nested data validation
Regex patterns for string matching
Status code verification
Response time thresholds
Custom assertion rules

`test_with_session`

Session-based authentication for complex flows:

Cookie jar management (JSESSIONID, session tokens)
CSRF token handling
Maintains session state across multiple requests
Perfect for testing authenticated user flows

`batch_test`

Run multiple API tests efficiently:

Parallel execution for performance
Sequential execution for dependent tests
Aggregated results and statistics
Ideal for regression testing and health checks

`analyze_api_spec`

Extract and understand OpenAPI/Swagger specifications:

Parse OpenAPI 2.0 and 3.0 specs
List all available endpoints
Understand request/response schemas
Generate sample requests automatically

`generate_test_scenarios`

AI-powered test generation from specs:

Creates comprehensive test suites from OpenAPI docs
Generates edge cases and boundary tests
Includes positive and negative test cases
Exports tests in various formats

Advanced Features

Smart Assertions: Validate responses with JSONPath, regex, and custom rules
Environment Management: Switch between dev/staging/prod seamlessly
Performance Profiling: Track response times and identify bottlenecks
Test Generation: Create tests from OpenAPI specs automatically
MCP Discovery: Find and test other MCP servers

Configuration Options

API Lab MCP works out of the box with zero configuration. You can customize behavior through tool parameters:

Tool-level Options

timeout: Request timeout in milliseconds (default: 30000) - available in all test tools
headers: Custom headers for requests - supports any HTTP headers
method: HTTP methods - GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS
body: Request payload - JSON, form data, or raw text

Environment Variables (HTTP Server Mode)

PORT: Server port for HTTP mode (default: 3001)
LOG_LEVEL: Logging verbosity - ERROR, WARN, INFO, DEBUG (default: INFO)

Common Issues & Solutions

Installation Issues

If npm install fails, try clearing npm cache: npm cache clean --force
On Windows, use npx.cmd instead of npx in the config

Connection Issues

Ensure the MCP server is listed in Claude's interface (look for the 🔌 icon)
Restart Claude Desktop after configuration changes
Check firewall settings for local development

Testing Issues

For large responses or slow APIs, increase timeout parameter in the tool call
Use session tools (test_with_session) for stateful API testing
For debugging, check Claude's developer console for detailed error messages

Full API Documentation →

🤝 Contributing

We love contributions! Whether it's a bug report, feature request, or pull request, all contributions are welcome.

# Fork and clone
git clone https://github.com/YOUR_USERNAME/api-lab-mcp.git

# Create your feature branch
git checkout -b feature/amazing-feature

# Make your changes and commit
git commit -m 'Add amazing feature'

# Push and create PR
git push origin feature/amazing-feature

See CONTRIBUTING.md for detailed guidelines.

🌟 Success Stories

"API Lab MCP reduced our API testing time by 70%. We now test as we code, right in Claude." - Senior Developer at TechCorp

"The natural language interface is a game-changer. Junior developers can now write complex API tests." - CTO at StartupXYZ

"We replaced three different tools with just API Lab MCP. It's that powerful." - DevOps Lead

📈 Roadmap

🛟 Support

📖 Documentation
🐛 Issue Tracker

📄 License

If API Lab MCP helps your development workflow, please consider giving it a ⭐

Built with ❤️ for developers who value their time

Name		Name	Last commit message	Last commit date
Latest commit History 13 Commits
.claude		.claude
docs		docs
examples		examples
src		src
tests		tests
.gitignore		.gitignore
.npmignore		.npmignore
.prettierrc		.prettierrc
README.md		README.md
TODO.md		TODO.md
claude_desktop_config.json		claude_desktop_config.json
claude_desktop_config_example.json		claude_desktop_config_example.json
jest.config.js		jest.config.js
package.json		package.json
plan.md		plan.md
test-auth.js		test-auth.js
test-mcp-features.sh		test-mcp-features.sh
test-mcp-http.sh		test-mcp-http.sh
tsconfig.json		tsconfig.json

atototo/api-lab-mcp

Folders and files

Latest commit

History

Repository files navigation