R2R

R2R: The first open source agentic retrieval system available via API that autonomously researches your documents, searches and scrapes the web, builds knowledge graphs, and delivers comprehensive, citation-backed answers through simple RESTful endpoints.

R2R is an advanced open-source AI retrieval system designed to streamline deep research and data analysis through a RESTful API. It autonomously processes documents, scrapes the web, and builds knowledge graphs to deliver citation-backed answers. Supporting Retrieval-Augmented Generation (RAG), R2R offers multimodal content ingestion, hybrid search, and agentic reasoning for complex queries. The system includes a Deep Research API for multi-step reasoning, fetching relevant data from both local knowledge bases and the internet. Users can integrate R2R via Python or JavaScript SDKs, with options for cloud deployment through SciPhi or self-hosting. Features like automatic entity extraction, user management, and hybrid search make it a versatile tool for research-intensive applications. A managed cloud option with a free tier is available, simplifying access without requiring a credit card. R2R is ideal for developers and researchers seeking scalable, AI-powered retrieval solutions.