Backend Documentation

This section contains documentation for the Python backend components of InsightHub.

Core Components

🎯 AI Processing Pipeline (Task 31 ✅)

• AI_PIPELINE.md - Complete orchestrator implementation
• LangGraph StateGraph topology
• Processing nodes (ContentFetcher, Summarizer, Embedding, ContentScorer, Storage)
• ErrorHandlerNode with classification & recovery
• Retry management & circuit breaker patterns
• Optimization pipeline with metrics-driven tuning

🔧 Infrastructure & Monitoring

• LANGSMITH_INTEGRATION.md - LangSmith monitoring setup
• LANGSMITH_OPERATIONS.md - Operational dashboards & workflows
• SUPABASE_INTEGRATION.md - Database integration

📊 API & Reference

• API_REFERENCE.md - Backend API endpoints

Architecture Overview

The backend follows a LangGraph-based orchestration pattern where content flows through specialized processing nodes:

START → ContentFetcher → Summarizer → Embedding → ContentScorer → Storage → END
                                     ↓ (on error)
                              ErrorHandlerNode

See AI_PIPELINE.md for complete technical details.