SahinDa/RAGify
Production-grade hybrid RAG backend built with FastAPI, ChromaDB Cloud, and Postgres (Neon) full-text search. Features structure-aware chunking, cross-encoder re-ranking via RRF, in-process conversation memory with rolling summaries, and a dual-mode runtime (zero-cost local DEV vs. Gemini-powered PROD).
GitHub repository with 5 stars and 1 forks.
Language: Python
Topics: chromadb, cross-encoder, fastapi, gemini-api, hybrid-search, llm, nvidia-nim, postgresql, rag, reranking