ModelRefs / Naive RAG Stack — Architecture Pattern
Naive RAG Stack — Architecture Pattern
Document ingestion → embeddings → vector store → top-k retrieval → prompt assembly → LLM generation. Minimal moving parts, maximum portability.
Overview
The canonical single-shot RAG topology: chunk documents, embed them, store in a vector DB, retrieve top-k at query time, stuff the context into a prompt, generate. Minimal moving parts, maximum portability.
When to use it: You need to ground an LLM in your documents with the smallest viable system.
Pattern details
| Pattern class | rag |
|---|---|
| Difficulty | beginner |
| Topology | pipeline |
| Also known as | basic rag, stuff rag, v1 rag |
| Last reviewed | 2026-06-07 |
Known failure modes
- Low recall — Retriever misses the relevant chunk. Mitigation: Add reranker or hybrid BM25+vector.
- Context overflow — Stuffed context exceeds the model window. Mitigation: Cap k; summarize chunks; switch to advanced RAG.
When not to use it
- Treating embeddings as the only signal — ignore lexical match.
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Naive RAG Stack — Architecture Pattern.
Frequently asked questions
When should I adopt the Naive RAG Stack?
You need to ground an LLM in your documents with the smallest viable system.
What are common failure modes of Naive RAG Stack?
Low recall • Context overflow
Is Naive RAG Stack production-ready?
Yes when paired with the safety controls and observability hooks documented on the pattern page.