
AI / Backend SystemAgentic Conversational RAG Backend
MemoryForge
A modular conversational RAG system orchestrated with LangGraph, combining BM25 and vector similarity search with Redis-backed memory across 100+ sessions. Integrated LLMLingua for 4–6× token compression with real-time streaming responses, deployed via Docker Compose with full observability.
LangGraphRedisFastAPIDockerBM25LLMLingua
Open Project