#3939th of 39 in RAG, knowledge and search
r2r
RAG server with REST API, hybrid search, knowledge graphs and a research agent
- Stars
- 8k
- License
- MIT
- Last commit
- Nov 2025
- Last release
- Jun 2025
- Language
- Python
7,663 on 1 Apr 20268,036 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
R2R is a retrieval-augmented generation server built around a RESTful API, with Python and JavaScript SDKs. It ingests documents (txt, pdf, json, png, mp3), runs hybrid semantic and keyword search with reciprocal rank fusion, and extracts entities and relationships into knowledge graphs. A Deep Research agent does multi-step reasoning over your knowledge base and optionally the web.
Who it is for: Developers building document Q&A or agentic RAG backends over a REST API
Strengths
- Hybrid search combines semantic and keyword results with reciprocal rank fusion
- Python and JavaScript SDKs over one REST API (examples use port 7272)
- Includes user auth, collections and document management
- Light mode runs from pip install; full mode uses Docker Compose with Postgres
Weaknesses
- Last release was 2025-06-06; later commits are unreleased
- README gives no RAM or hardware requirements
- Examples assume hosted LLMs (OpenAI, Anthropic); local model setup not shown
- No bundled chat UI described in the README
What it needs
- Compose
- Needs OpenAI API key (default in quick start), Postgres (full mode via Docker Compose)
- Models: OpenAI, Anthropic
- port 7272
Also in RAG, knowledge and search
See all 39| Rank | Project | Score |
|---|---|---|
| 1 | deeptutorSelf-hosted AI tutor with knowledge bases, guided learning, books and quizzes | 82 out of 100 |
| 2 | LightRAGGraph-based RAG framework with a REST API server and web UI | 78 out of 100 |
| 3 | MorphicAI answer engine with cited search results and generative UI components | 77 out of 100 |
| #4 | MaxKBEnterprise knowledge-base agent platform with RAG, workflows and MCP tools | 75 out of 100 |
| #5 | RAGFlowRAG engine with deep document parsing, template chunking and agentic retrieval | 74 out of 100 |