Debugg AI MCP

What is Debugg AI MCP about?

Debugg AI MCP provides an AI‑powered toolkit that lets developers describe tests in plain language, run them instantly in a real browser, monitor console output, network traffic, and screenshots live, and organize results into test suites that can be viewed from the Debugg.AI dashboard.

How to use Debugg AI MCP?

1. Obtain an API key from https://debugg.ai.
2. Run the server via NPX (recommended):

   npx -y @debugg-ai/debugg-ai-mcp
   

   or with Docker if preferred.
3. Send MCP requests from any MCP‑compatible client (Claude Desktop, LangChain, etc.) using the provided tools such as debugg_ai_test_page_changes, debugg_ai_create_test_suite, debugg_ai_start_live_session, etc.
4. Monitor results in real time via the Debugg.AI web dashboard or retrieve them programmatically with the status tools.

Key Features of Debugg AI MCP

• Zero‑config E2E testing – write tests like “Test the user login flow on my app running on port 3000”.
• Live session monitoring – real‑time console logs, network traffic, and screenshots during test execution.
• Test suite management – create, list, and track both feature‑based and commit‑based test suites.
• Universal compatibility – works with any MCP‑compatible client, no SDK lock‑in.
• CI/CD integration – results automatically appear in the Debugg.AI dashboard for visibility in pipelines.
• Docker and NPX deployment – simple one‑liner start up.

Use Cases of Debugg AI MCP

• Rapid prototyping: QA engineers can instantly validate new UI flows without writing Selenium or Playwright scripts.
• Commit validation: Generate and run tests automatically for every git commit to catch regressions early.
• Live debugging: Observe browser behavior while a test runs to diagnose flaky failures.
• Continuous testing: Embed the MCP server in CI pipelines to provide centralized test reporting.
• Cross‑team collaboration: Non‑technical stakeholders describe expected behavior in plain language, and the AI creates executable tests.

FAQ from Debugg AI MCP

Q: Do I need to install any browser drivers? A: No. The server runs headless Chromium internally, managed by the AI service.

Q: Which programming languages are supported for test definitions? A: Tests are defined via natural‑language prompts; the underlying implementation uses Node.js/Playwright, but you never write code directly.

Q: Can I run the server locally without an internet connection? A: A valid Debugg.AI API key is required because the AI engine runs in the cloud. The server itself can be hosted locally.

Q: How are test results stored? A: Results are sent to the Debugg.AI platform and can also be retrieved via the debugg_ai_get_test_status tool.

Q: Is there a limit on the number of concurrent live sessions? A: Limits are governed by your Debugg.AI subscription tier.

🧪 Official MCP Server for Debugg AI

AI-powered development and testing toolkit implementing the Model Context Protocol (MCP), designed to give AI agents comprehensive testing, debugging, and code analysis capabilities.

Transform your development workflow with:

• Zero-config E2E testing - Run browser tests with natural language descriptions
• Live session monitoring - Real-time browser console, network, and screenshot monitoring
• Test suite management - Create and manage comprehensive test suites
• Seamless CI/CD integration - View all test results in your Debugg.AI App dashboard

---

🚀 Features

12 Focused Development Tools

• 🧪 E2E Testing Suite - Run browser tests, create test suites, and generate commit-based tests
• 🖥️ Live Session Monitoring - Real-time browser console, network traffic, and screenshot monitoring
• 📊 Test Management - List, create, and track test suites and commit-based test suites
• 📱 Real-time Progress - Live updates with screenshots and step-by-step execution
• 🌐 Universal Compatibility - Works with any MCP-compatible client (Claude Desktop, LangChain, etc.)

---

Examples

Input prompt: "Test the ability to create an account and login"

Results:

**Task Completed**

- Duration: 86.80 seconds
- Final Result: Successfully completed the task of signing up and logging into the account with the email 'alice.wonderland1234@example.com'.
- Status: Success

Full Demo:

Watch a more in-depth, Full Use Case Demo

---

🛠️ Quick Setup

1. Get Your API Key

Create a free account at debugg.ai and generate your API key.

2. Choose Your Installation Method

Option A: NPX (Recommended)

npx -y @debugg-ai/debugg-ai-mcp

Option B: Docker

docker run -i --rm --init \
  -e DEBUGGAI_API_KEY=your_api_key \
  quinnosha/debugg-ai-mcp

---

🧰 Available Tools

E2E Testing Tools

• debugg_ai_test_page_changes - Run browser tests with natural language descriptions
• debugg_ai_create_test_suite - Create organized test suites for features
• debugg_ai_create_commit_suite - Generate tests based on git commits
• debugg_ai_get_test_status - Monitor test execution and results

Test Management Tools

• debugg_ai_list_tests - List all E2E tests with filtering and pagination
• debugg_ai_list_test_suites - List all test suites with filtering options
• debugg_ai_list_commit_suites - List all commit-based test suites

Live Session Monitoring Tools

• debugg_ai_start_live_session - Start a live browser session with real-time monitoring
• debugg_ai_stop_live_session - Stop an active live session
• debugg_ai_get_live_session_status - Get the current status of a live session
• debugg_ai_get_live_session_logs - Retrieve console and network logs from a live session
• debugg_ai_get_live_session_screenshot - Capture screenshots from an active live session

---

⚙️ Configuration

For Claude Desktop

Add this to your MCP settings file:

{
  "mcpServers": {
    "debugg-ai-mcp": {
      "command": "npx",
      "args": ["-y", "@debugg-ai/debugg-ai-mcp"],
      "env": {
        "DEBUGGAI_API_KEY": "your_api_key_here"
      }
    }
  }
}

Optional Environment Variables

# Required
DEBUGGAI_API_KEY=your_api_key

# Optional (with sensible defaults)
DEBUGGAI_LOCAL_PORT=3000                    # Your app's port
DEBUGGAI_LOCAL_REPO_NAME=your-org/repo      # GitHub repo name
DEBUGGAI_LOCAL_REPO_PATH=/path/to/project   # Project directory

💡 Usage Examples

Run a Quick E2E Test

"Test the user login flow on my app running on port 3000"

Analyze Your Project

"What frameworks and languages are used in my codebase?"

Get Issue Insights

"Show me all high-priority issues in my project"

Generate Test Coverage

"Generate test coverage for the authentication module"

---

🧑‍💻 Local Development

# Install dependencies
npm install

# Run tests
npm test

# Build project
npm run build

# Start server locally
node dist/index.js

---

📁 Project Structure

debugg-ai-mcp/
├── config/          # Configuration management  
├── tools/           # 14 MCP tool definitions
├── handlers/        # Tool implementation logic
├── services/        # DebuggAI API integration
├── utils/           # Shared utilities & logging
├── types/           # TypeScript type definitions
├── __tests__/       # Comprehensive test suite
└── index.ts         # Main server entry point

---

🚀 Publishing & Releases

This project uses automated publishing to NPM. Here's how it works:

Automatic Publishing

• Every push to main triggers automatic NPM publishing
• Only publishes if the version doesn't already exist
• Includes full test suite validation and build verification

Version Management

# Bump version locally
npm run version:patch  # 1.0.15 → 1.0.16
npm run version:minor  # 1.0.15 → 1.1.0
npm run version:major  # 1.0.15 → 2.0.0

# Check package contents
npm run publish:check

Manual Version Bump via GitHub

1. Go to Actions → Version Bump
2. Click "Run workflow"
3. Select version type or enter custom version
4. Workflow will update version and trigger publish

Setup for Contributors

See .github/PUBLISHING_SETUP.md for complete setup instructions.

---

💬 Support & Links

• 📖 Documentation: debugg.ai/docs
• 🐛 Issues: GitHub Issues
• 💬 Discord: Join our community
• 🌐 Dashboard: app.debugg.ai

---

🔒 License

Apache-2.0 License © 2025 DebuggAI

---