R2R

Why we included this project

R2R packages retrieval-augmented generation so you don't have to assemble a parser, an index, and an orchestration layer yourself. It combines document ingestion, hybrid semantic and keyword search, automatic knowledge-graph extraction, and an agentic reasoning layer behind one RESTful API, so raw files become a queryable knowledge base in a single service. Python and JavaScript SDKs let you call search, RAG with citations, and a deep-research agent from an existing app, and Docker-based self-hosting keeps your data on your own infrastructure. It also ships with user and access management and a document collection system, which makes it a reasonable backend for a small team that wants room to grow rather than a demo notebook.