yacine4700/easy-3olom-backend

Backend for Easy 3olom — an AI-powered educational assistant for Algerian students. # Easy 3olom Backend Backend for Easy 3olom, an AI-powered educational assistant designed primarily for Algerian high-school and Baccalaureate students. ## Overview Easy 3olom aims to provide students with an educational AI assistant through Telegram. The system will support: - Text-based questions - Voice messages - AI-powered explanations - Curriculum-aware RAG - Educational methodology - Conversation context - Subscription management - Usage limits - Redis caching - Supabase/PostgreSQL - Multiple LLM providers ## Architecture The backend will use a modular monolith architecture. Main technologies: - Python - FastAPI - Redis - Supabase/PostgreSQL - Telegram Bot API - LLM APIs - RAG - Docker n8n may be used for secondary automation, but it must not be required for the core operation of the application. ## Development Philosophy The project should prioritize: - Simplicity - Reliability - Maintainability - Low operating cost - Good performance - Incremental development - Avoiding unnecessary abstractions and microservices The initial deployment target is a single Hetzner server using Docker Compose. ## Development Status The project is currently being rebuilt from an existing n8n prototype into a dedicated Python/FastAPI backend. The backend will be implemented incrementally, starting with the Telegram webhook and core infrastructure.