All open source
Open SourceRAG7k

R2R

Agentic RAG with a REST API: multimodal ingestion and GraphRAG.

R2R (Reason to Retrieve) is an open-source, production-oriented RAG system exposed through a RESTful API. It combines multimodal document ingestion, hybrid semantic and keyword search, automatic knowledge-graph extraction, and user-level access controls, plus an agentic Deep Research API for multi-step reasoning queries. It is notable for packaging retrieval, graph construction, and document/permission management behind a deployable API.

Repository

SciPhi-AI/R2R

Language

Python

What you'd build with it

  • Deploying a RAG backend with authentication and per-user document access controls behind a single API
  • Combining vector, keyword, and knowledge-graph retrieval for context-aware question answering
  • Running multi-step agentic research queries over an ingested document corpus

Tags

raggraphragapi