Ad Intelligence Extractor

<h1>Ad Intelligence Extractor</h1> <h2>Overview</h2> <p>A full-stack competitive intelligence platform that automates the extraction and analysis of competitor advertising across four major ad transparency libraries. Built with Playwright browser automation as the primary extraction method, with official API fallbacks for reliability.</p> <h2>Key Features</h2> <h3>Multi-Platform Extraction</h3> <p>Unified extraction from Meta Ad Library, TikTok Creative Center, Google Ads Transparency Center, and LinkedIn Ad Library. Each platform uses specialised DOM selectors with retry logic and rate limiting for reliable data collection.</p> <h3>AI-Powered Analysis</h3> <p>Integrates with Claude and OpenAI to deliver strategic competitive insights beyond raw data. The AI pipeline includes ad copy analysis, campaign pattern detection, audience insights, and competitive battle analysis.</p> <h3>Intelligent Categorisation</h3> <p>Automatically classifies extracted ads into 11 marketing categories including testimonial, offer/promo, educational, product feature, brand awareness, urgency/scarcity, problem/solution, and comparison using rule-based confidence scoring.</p> <h3>Landing Page Intelligence</h3> <p>Analyses destination URLs to extract conversion elements including CTAs, form friction scores, trust signals, and offer structures for a complete view of the competitor funnel.</p> <h3>Multi-Format Reporting</h3> <p>Generates comprehensive reports in Markdown, PowerPoint, Word, Excel, PDF, and JSON formats suitable for client presentations, team collaboration, and programmatic access.</p> <h3>Real-Time Progress</h3> <p>WebSocket-based live updates during extraction jobs, with a browser-based dashboard for non-technical users alongside a full CLI for advanced usage.</p> <h3>MCP Integration</h3> <p>Model Context Protocol server enables natural language queries about competitor ads directly through Claude Desktop.</p> <h2>Technical Architecture</h2> <p>The extraction strategy prioritises Playwright browser automation for full data access and screenshot capture, falling back to official APIs when scraping encounters rate limits. The Express.js backend handles job orchestration, while the WebSocket layer provides real-time progress to the React-based dashboard.</p> <h2>Use Cases</h2> <ul> <li><strong>Marketing Teams</strong>: Competitive strategy research, creative inspiration, market gap identification</li> <li><strong>Agencies</strong>: Client onboarding analysis, data-driven strategy, multi-client comparison</li> <li><strong>Business Development</strong>: Prospect research, industry trend analysis, market intelligence</li> </ul>