29.9K ·GitHub
What is this?
A community-driven hub of advanced techniques for Retrieval-Augmented Generation (RAG) — grounding LLM answers in retrieved documents, distilled into runnable notebook tutorials.
A community-driven hub of advanced techniques for Retrieval-Augmented Generation (RAG) — grounding LLM answers in retrieved documents, distilled into runnable notebook tutorials.Is this an installable library, or a set of tutorials?A set of tutorials, not a package: 42+ self-contained Jupyter notebooks, each teaching one RAG technique with intuition, code and references.
A set of tutorials, not a package: 42+ self-contained Jupyter notebooks, each teaching one RAG technique with intuition, code and references.How do I work through a technique?Clone the repo, open the technique's notebook in Jupyter and run it top to bottom — each notebook is a standalone lesson you can adapt to your own stack.
Clone the repo, open the technique's notebook in Jupyter and run it top to bottom — each notebook is a standalone lesson you can adapt to your own stack.What do I need installed and configured to run the notebooks?Python with Jupyter. Per the repo topics, the notebooks are built on LangChain and LlamaIndex and use OpenAI's GPT models, so plan on an OpenAI API key for most of them.
Python with Jupyter. Per the repo topics, the notebooks are built on LangChain and LlamaIndex and use OpenAI's GPT models, so plan on an OpenAI API key for most of them.Any gotchas before I dive in?It teaches techniques rather than shipping a production library — you port the code into your own system. Notebooks also call hosted LLM APIs, so running them all consumes API credits.
It teaches techniques rather than shipping a production library — you port the code into your own system. Notebooks also call hosted LLM APIs, so running them all consumes API credits.Will a RAG beginner benefit, or is it for people already building systems?Both: material runs from foundational basics to cutting-edge, and it is written for researchers and practitioners pushing what RAG can do.
Both: material runs from foundational basics to cutting-edge, and it is written for researchers and practitioners pushing what RAG can do.Can I run this on a regular laptop, or do I need special hardware?Any machine with Python and Jupyter works — Windows, macOS or Linux; the LLM work happens through cloud APIs, not your hardware.
Any machine with Python and Jupyter works — Windows, macOS or Linux; the LLM work happens through cloud APIs, not your hardware.What kinds of techniques does the collection cover?Everything from Basic RAG, chunk-size optimization and proposition chunking to query transformations, HyDE (hypothetical document embeddings) and agentic RAG, plus recent additions like MemoRAG and end-to-end RAG evaluation.
Everything from Basic RAG, chunk-size optimization and proposition chunking to query transformations, HyDE (hypothetical document embeddings) and agentic RAG, plus recent additions like MemoRAG and end-to-end RAG evaluation.Where should a beginner start?In the Foundational section: Basic RAG first, then RAG with CSV Files, Reliable RAG, Optimizing Chunk Sizes and Proposition Chunking, before moving to Query Enhancement techniques like HyDE.
In the Foundational section: Basic RAG first, then RAG with CSV Files, Reliable RAG, Optimizing Chunk Sizes and Proposition Chunking, before moving to Query Enhancement techniques like HyDE.Is the collection still growing, or is it stale?Alive and growing: the README lists recent additions (MemoRAG, End-to-End RAG Evaluation, Open-RAG-Eval, JSON RAG), counts 42 notebooks and promises regular updates. The repo is not archived and has about 29.8k stars.
Alive and growing: the README lists recent additions (MemoRAG, End-to-End RAG Evaluation, Open-RAG-Eval, JSON RAG), counts 42 notebooks and promises regular updates. The repo is not archived and has about 29.8k stars.Is everything here free?The notebooks are free. The author monetizes companion material — a 20-module Prompt to Production course and the 400-page book RAG Made Simple — both linked from the README.
The notebooks are free. The author monetizes companion material — a 20-module Prompt to Production course and the 400-page book RAG Made Simple — both linked from the README.What are similar projects?agentic-rag-for-dummies, generative-ai-for-beginners, llm-universe, llm-cookbook, LEANN, GenAI_Agents, Hands-On-Large-Language-Models, ai-system-design
agentic-rag-for-dummies, generative-ai-for-beginners, llm-universe, llm-cookbook, LEANN, GenAI_Agents, Hands-On-Large-Language-Models, ai-system-designWhich topics is it in?llama-index, agentic-rag, tutorials, retrieval-augmented-generation, llms, langchain, vector-database, semantic-search, embeddings, gpt, course, evaluation
llama-index, agentic-rag, tutorials, retrieval-augmented-generation, llms, langchain, vector-database, semantic-search, embeddings, gpt, course, evaluation