Skip to main content

MCP Server Setup

Configure the MCP server for AI agent integration.

Prerequisites​

Requirements​

  • Node.js: Version 16+
  • DebuggAI Account: Sign up
  • API Key: From DebuggAI settings

Get API Key​

  1. Sign in at debugg.ai
  2. Settings → API Keys
  3. Generate new key
  4. Copy and save securely

Installation Methods​

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

NPM Global​

npm install -g @debugg-ai/debugg-ai-mcp
debugg-ai-mcp

Docker​

docker run -e DEBUGGAI_API_KEY=your_key debugg-ai/mcp-server

Claude Desktop Integration​

1. Locate Config File​

macOS:

~/Library/Application Support/Claude/claude_desktop_config.json

Windows:

%APPDATA%/Claude/claude_desktop_config.json

Linux:

~/.config/Claude/claude_desktop_config.json

2. Add Configuration​

claude_desktop_config.json
{
  "mcpServers": {
    "debugg-ai-mcp": {
      "command": "npx",
      "args": ["-y", "@debugg-ai/debugg-ai-mcp"],
      "env": {
        "DEBUGGAI_API_KEY": "your_api_key_here",
        "DEBUGGAI_LOCAL_PORT": "3000"
      }
    }
  }
}

3. Restart Claude Desktop​

Close and reopen Claude Desktop to load the MCP server.

Environment Variables​

Required Variables​

DEBUGGAI_API_KEY=your_key          # API key from debugg.ai
DEBUGGAI_LOCAL_PORT=3000           # Your app's port

Optional Variables​

# Test credentials
TEST_USERNAME_EMAIL=test@example.com
TEST_USER_PASSWORD=testpassword123

# Project context
REPO_NAME=my-project
BRANCH_NAME=main
REPO_PATH=/path/to/repo

# Advanced settings (rarely needed)
DEBUGGAI_TIMEOUT=30000             # Test timeout in milliseconds
DEBUGGAI_RETRIES=2                 # Number of retry attempts

Docker Deployment​

Basic Docker​

Dockerfile
FROM node:18-alpine

WORKDIR /app

# Install MCP server
RUN npm install -g @debugg-ai/debugg-ai-mcp

# Expose port
EXPOSE 3000

# Start server
CMD ["debugg-ai-mcp"]

Build and run:

docker build -t debuggai-mcp .
docker run -d \
  -e DEBUGGAI_API_KEY=your_key \
  -e DEBUGGAI_LOCAL_PORT=3000 \
  -p 3000:3000 \
  debuggai-mcp

Docker Compose​

docker-compose.yml
version: '3.8'
services:
  debuggai-mcp:
    image: node:18-alpine
    command: sh -c "npm install -g @debugg-ai/debugg-ai-mcp && debugg-ai-mcp"
    environment:
      - DEBUGGAI_API_KEY=your_key
      - DEBUGGAI_LOCAL_PORT=3000
    ports:
      - "3000:3000"

Testing Setup​

1. Verify Installation​

# Check if MCP server is available
npx @debugg-ai/debugg-ai-mcp --version

2. Test in Claude Desktop​

In Claude Desktop, type:

Test my homepage loads correctly using debugg_ai_test_page_changes

3. Check Logs​

Monitor the MCP server logs for connection and test execution.

Advanced Configuration​

Custom MCP Server​

custom-mcp-server.js
const { MCPServer } = require('@debugg-ai/debugg-ai-mcp');

const server = new MCPServer({
  apiKey: process.env.DEBUGGAI_API_KEY,
  localPort: 3000,
  timeout: 30000,
  retries: 2
});

server.start();

Programmatic Usage​

example.js
const { DebuggAIClient } = require('@debugg-ai/debugg-ai-mcp');

const client = new DebuggAIClient({
  apiKey: process.env.DEBUGGAI_API_KEY
});

async function runTest() {
  const result = await client.testPageChanges({
    description: 'Test user registration form',
    localPort: 3000
  });
  
  console.log('Test result:', result);
}

runTest();

Troubleshooting​

MCP Server Not Starting​

# Check Node.js version
node --version

# Verify API key
echo $DEBUGGAI_API_KEY

# Test connection
curl -I localhost:3000

Claude Desktop Issues​

  1. Restart Claude: Close and reopen application
  2. Check JSON syntax: Validate claude_desktop_config.json format
  3. Check logs: Look for MCP connection errors in Claude Desktop

Connection Problems​

  1. Firewall: Allow localhost connections
  2. Port conflicts: Use different port
  3. Network: Verify internet connectivity

API Issues​

  1. Invalid key: Regenerate in DebuggAI settings
  2. Quota exceeded: Check usage limits
  3. Account suspended: Contact support

Production Deployment​

Security Considerations​

  • Store API keys securely (env variables, secrets)
  • Use HTTPS for external connections
  • Limit network access to required ports
  • Monitor for unusual activity

Scaling​

  • Use load balancers for multiple instances
  • Implement health checks
  • Set up monitoring and alerting
  • Configure auto-restart on failures

Monitoring​

# Check server status
curl -f localhost:3000/health || echo "Server down"

# Monitor logs
tail -f /var/log/debuggai-mcp.log

# Resource usage
docker stats debuggai-mcp