R2R

// Overview

R2R (Reason to Retrieve) is an open-source retrieval system for building Retrieval-Augmented Generation (RAG) applications. It runs as a server with a RESTful API and handles the parts of a RAG pipeline you would otherwise wire together yourself: ingesting documents, searching over them, and generating answers with citations.

It is aimed at developers who need more than a single-file demo. Alongside basic search and RAG, it includes multimodal ingestion, hybrid (semantic plus keyword) search, automatic knowledge-graph extraction, user and access management, and an agentic Deep Research API that runs multi-step reasoning over your knowledge base and the web.

As a RAG framework, R2R sits between raw vector stores and your application. You talk to it through Python or JavaScript SDK clients (or the API directly), so the retrieval and document-management logic lives in one service rather than scattered across your app code.

// What it does

• Multimodal ingestion: parse .txt, .pdf, .json, .png, .mp3, and more file types
• Hybrid search combining semantic and keyword retrieval with reciprocal rank fusion
• Automatic knowledge graphs with entity and relationship extraction
• Agentic RAG, including a Deep Research API for multi-step reasoning over your data and the internet
• Built-in user and access management with authentication and collections
• RESTful API with Python (r2r) and JavaScript (r2r-js) SDK clients

// Getting started

Install R2R, start the server in light mode, then connect a client to ingest documents and run searches.

pip install r2r
export OPENAI_API_KEY=sk-...
python -m r2r.serve

pip install r2r  # Python
# or
npm i r2r-js    # JavaScript

from r2r import R2RClient
client = R2RClient(base_url="http://localhost:7272")

# Ingest your own document
client.documents.create(file_path="/path/to/file")

# Basic search
results = client.retrieval.search(query="What is DeepSeek R1?")

# RAG with citations
response = client.retrieval.rag(query="What is DeepSeek R1?")

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Add a searchable knowledge base with cited answers to an internal app without building the RAG pipeline from scratch
• Ingest mixed file types (PDFs, images, audio, JSON) and query them through one API
• Build a knowledge graph from a document collection to surface entities and relationships
• Run multi-step Deep Research queries that reason over both your data and the web

// How R2R compares

R2R alongside other open-source rag frameworks & platforms tools AI/TLDR tracks, ranked by GitHub stars.