Browser Explorer is an AI-powered web browsing agent that automatically explores websites, tests interactive elements, and generates comprehensive Playwright test suites. Built with a modular architecture combining traditional crawling with AI-enhanced capabilities, it's designed to be the ultimate tool for automated web testing and exploration.

• 🔍 Intelligent Web Crawling: Breadth-first search algorithm for systematic website exploration
• 🤖 AI-Powered Element Detection: Hybrid AI + traditional detection for 25+ element types
• 🎯 Smart Interaction: Automated interaction with forms, buttons, dropdowns, modals, and more
• 📸 Comprehensive Recording: Complete user path capture with screenshots and timing
• 🧪 Test Generation: Automatic Playwright/Cypress/Puppeteer test suite generation
• 📄 Page Object Model: Clean, maintainable test structure with POM pattern

• 🔐 Multi-Strategy Authentication: Support for Basic, OAuth, MFA, and API key auth
• 🥷 Stealth Mode: Advanced anti-bot detection evasion
• 🧩 CAPTCHA Handling: Detection and solving capabilities
• 📊 Distributed Crawling: Redis-based queue for scalable operations
• 🐳 Docker Support: Full containerization with docker-compose
• 📈 Monitoring & Metrics: Built-in observability and health checks

• Complete infrastructure and crawling engine
• Traditional element detection and interaction
• Test generation pipeline with multiple frameworks
• Authentication system and session management
• CLI interface with comprehensive commands
• Docker containerization and deployment ready

• Mastra AI agent integration (shell classes only)
• Stagehand browser automation tools (not integrated)
• Natural language capabilities
• Self-healing test features

Note: While the infrastructure is production-ready, the AI capabilities that differentiate this from traditional crawlers are pending integration. See TODO.md for the complete roadmap.

• Node.js 20+
• Docker and Docker Compose (for full deployment)
• Redis (for distributed crawling)
• PostgreSQL (for data persistence)

# Clone the repository
git clone https://github.com/yourusername/browser-explorer.git
cd browser-explorer

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env

# Build the project
npm run build

# Run with Docker (recommended)
docker-compose up -d

# Or run locally
npm start

# Initialize configuration
browser-explorer init

# Crawl a website and generate tests
browser-explorer crawl https://example.com --output ./tests --depth 3

# Run specific test generation
browser-explorer test https://example.com/login --framework playwright

# Validate generated tests
browser-explorer validate ./tests

# Start monitoring dashboard
browser-explorer serve --port 3000

# Debug mode with verbose logging
browser-explorer debug https://example.com --verbose

import { BrowserExplorer } from 'browser-explorer';

// Initialize the explorer
const explorer = new BrowserExplorer({
  headless: true,
  depth: 3,
  parallel: 5,
  stealth: true
});

// Start exploration
const results = await explorer.explore('https://example.com', {
  authentication: {
    strategy: 'basic',
    credentials: { username: 'user', password: 'pass' }
  },
  output: {
    directory: './generated-tests',
    framework: 'playwright'
  }
});

// Generate Page Object Model
await explorer.generatePageObjects(results);

Browser Explorer
├── CLI/API Layer         # User interfaces
├── Core Services         # Orchestration and workflow
├── Components           # Element detection, interaction, recording
├── Infrastructure      # Config, logging, metrics, storage
└── Runtime            # Playwright, Redis, Docker, Node.js

• Crawler Module: BFS, resilient, and distributed crawling strategies
• Detection Module: AI-ready element detection with 25+ element types
• Generation Module: Multi-framework test generation with POM
• Authentication Module: Pluggable auth strategies and session management
• Monitoring Module: Metrics, traces, and health monitoring

See ARCHITECTURE.md for detailed architecture documentation.

crawler:
  depth: 3
  parallel: 5
  timeout: 30000
  respectRobotsTxt: true

detection:
  useAI: false  # Will be true when AI integration is complete
  fallbackToSelectors: true

generation:
  framework: playwright
  outputDir: ./tests
  generatePageObjects: true
  
browser:
  headless: true
  viewport:
    width: 1920
    height: 1080

# Database
DATABASE_URL=postgresql://user:pass@localhost:5432/browser_explorer

# Redis
REDIS_URL=redis://localhost:6379

# API Keys (when AI integration is complete)
OPENAI_API_KEY=your-key-here
ANTHROPIC_API_KEY=your-key-here

# Monitoring
ENABLE_METRICS=true
METRICS_PORT=9090

# Install dependencies
npm install

# Run in development mode
npm run dev

# Run tests
npm test

# Run linting
npm run lint

# Type checking
npm run typecheck

# Unit tests
npm run test:unit

# Integration tests
npm run test:integration

# Performance tests
npm run test:performance

# Self-test the system
npm run self-test

# Build and start all services
docker-compose up -d

# View logs
docker-compose logs -f

# Scale crawler workers
docker-compose up -d --scale crawler=5

# Stop all services
docker-compose down

The system includes comprehensive monitoring capabilities:

• Metrics: Application and infrastructure metrics
• Tracing: Distributed tracing for request flows
• Health Checks: Liveness, readiness, and dependency checks
• Dashboards: Built-in monitoring dashboard at /metrics

We welcome contributions! Please see our Contributing Guide for details.

1. Mastra AI Integration - Implement proper AI agent configurations
2. Stagehand Integration - Add natural language browser control
3. Self-Healing Tests - AI-powered test maintenance
4. Enhanced Detection - Natural language element queries

• Getting Started Guide - Complete setup and first steps
• Quick Examples - Working code examples
• Configuration Guide - Setup and configuration

• Documentation Index - Complete documentation overview
• API Reference - All classes, interfaces, and functions
• Architecture Overview - System design and patterns
• Implementation Status - Current gaps and roadmap

• Browser Agent - Core browser automation
• Authentication System - Multi-strategy auth
• AI Element Detection - Element identification
• Web Crawler - Systematic web exploration
• Test Generation - Automated test creation

• Contributing Guide - How to contribute
• Development Setup - Local development
• Testing Guide - Running and writing tests
• Docker Guide - Containerization

• AI Integration - Mastra and AI workflows
• Performance Optimization - Scaling and optimization
• Security Guide - Security best practices
• Troubleshooting - Common issues and solutions

• Server Mode: CLI advertises but not implemented (src/cli/BrowserExplorerCLI.ts:334)
• Authentication Setup: CLI auth command exists but non-functional (src/cli/BrowserExplorerCLI.ts:417)
• Stagehand AI Integration: Core AI detection completely disabled (src/detectors/AIElementDetector.ts:23-28)
• Test Generation: Placeholder implementations for core features (src/generation/TestGenerator.ts:497-507)

• OAuth authentication flows (framework exists, needs provider implementations)
• MFA support (basic structure, needs TOTP integration)
• Browser-Crawler integration (service architecture incomplete)
• CLI debug commands (advertised but not functional)

• Basic browser automation and control
• Username/password authentication
• Selector-based element detection
• Basic test generation templates
• Configuration management
• Session persistence
• Docker infrastructure

See Implementation Status Report for detailed analysis of all gaps, priorities, and planned fixes.

• Mastra agent implementation
• Stagehand tool integration
• Basic natural language support

• Self-healing test capabilities
• AI-powered visual regression
• Intelligent assertion generation

• Multi-agent collaboration
• Advanced learning capabilities
• Enterprise integrations

This project is licensed under the MIT License - see the LICENSE file for details.

• Built with Playwright for browser automation
• Designed for Mastra AI framework integration
• Prepared for Stagehand natural language control

Made with ❤️ by the Browser Explorer Team

Report Bug • Request Feature • Documentation