RAG Techniques is an open-source educational repository that demonstrates a broad range of retrieval-augmented generation methods. It is organized as practical examples showing how different retrieval, indexing, chunking, reranking, query, and context-management strategies can improve the quality of LLM applications grounded in external information.

Rather than presenting RAG as a single fixed architecture, the project explores multiple techniques that can be compared and combined depending on the data and task. The examples help developers understand tradeoffs between simple retrieval pipelines and more advanced approaches designed to improve relevance, reasoning, or answer quality.

The repository is useful for AI engineers, students, and researchers building document question answering, knowledge assistants, search systems, and agent memory. It functions both as a learning resource and as a cookbook of implementation patterns that can be adapted when a basic vector-search RAG pipeline is no longer sufficient.