Naive RAG

A linear, one-concept-per-lesson path through Naive RAG, the baseline retrieval-augmented generation architecture, built from scratch with no framework, from your first embedding to a complete FastAPI RAG service backed by ChromaDB. 26 lessons · 3 tiers.

Prerequisites: Comfortable writing basic Python (functions, imports, running a script). No LangChain or vector-database experience needed - this course builds naive RAG from raw API calls and a plain Python list, deliberately without a framework. This site's RAG Fundamentals course is a useful (not required) primer: it maps where naive RAG sits among the nine retrieval architectures, of which this course covers only the first.

Lessons use Google's Gemini free tier (gemini-embedding-001 for embeddings, gemini-3.5-flash-lite for chat), the same models as this site's other courses. No Docker and no database for most of the course - the vector store starts as a plain Python list; Lesson 20 onward swaps that for a local ChromaDB instance, still no server or account required.

Built on ChromaDB, the open-source vector database this course graduates to in the Advanced tier.