rag-zero-to-hero-guide: Learn Retrieval-Augmented Generation

Overview

rag-zero-to-hero-guide is an educational collection for learning retrieval-augmented generation (RAG), the approach of grounding language-model responses in retrieved information. It brings together introductory material, practical notebooks, evaluation lessons, tool references, and research papers in one repository.

It suits developers and learners who want a guided path from RAG fundamentals to topics such as agentic and graph-based RAG. The repository is a curriculum and reference collection rather than a single installable RAG application.

Key Features

• Conceptual lessons explain RAG basics, its benefits and challenges, key terminology, and a learning roadmap.
• Jupyter notebooks demonstrate RAG from scratch and implementations using LangChain.
• Example notebooks cover RAG over website content and YouTube transcripts, plus an agentic RAG example using CrewAI.
• Evaluation material introduces retriever and generator metrics and includes RAGAS, DeepEval, and hallucination-detection notebooks.
• A categorized toolkit directory points to frameworks, data extraction tools, vector databases, chunking and reranking tools, agentic and graph RAG projects, and evaluation libraries.
• A survey-paper index links to research on general RAG, graph RAG, agentic RAG, evaluation, and best practices.

Use Cases

• A developer new to RAG can follow the concepts and roadmap before choosing an implementation framework.
• A learner can adapt the notebooks to experiment with retrieval over documents, web pages, or video transcripts.
• An engineer developing a RAG system can consult the evaluation lessons when comparing retrieval and generation quality.
• A researcher or technical lead exploring advanced approaches can use the toolkit and survey-paper links to find relevant starting points.

Getting Started

Start with the RAG Basics Course and open a notebook or lesson that matches your goal. See the repository README for the full course index and links.

Alternatives

• rag-from-scratch: It teaches RAG through a focused notebook progression on indexing, retrieval, and generation, while rag-zero-to-hero-guide also covers evaluation and advanced approaches.

Considerations

• This is a collection of learning materials and external tool references, not a packaged application with a single setup command.
• The notebooks use different frameworks and libraries, so dependencies and any model or service requirements depend on the example. Check each notebook before running it.
• The README provides links and topic descriptions, but the repository input does not establish how complete or recently updated each linked resource is.