Guides
Longer explainers on local, private RAG: why retrieval fails, which retrieval method fits, and what hardware to buy. For tool-by-tool decisions, see the comparisons. For step-by-step setups, see the recipes.
All guides
- Why your RAG retrieval fails (and how to test it)
Common local RAG retrieval failures (chunking, embeddings, top-k, rerankers, filters, PDF parsing) and a small golden-passage eval. - Vector vs. vectorless RAG: when a ToC tree beats embeddings
Why similarity is not relevance, how table-of-contents tree retrieval works, and which local tools use which retrieval. - LangChain RAG vs. local RAG: which should you use?
Code framework or ready-made offline app: who each is for, setup, cost, retrieval and evals. - Local AI on a small box or under $2,500
Mini-PC, Mac mini and used-GPU options for local RAG, with links to tools and the cost calculator.
Head-to-head comparisons
- AnythingLLM vs Open WebUI
- AnythingLLM vs RAGFlow
- AnythingLLM vs PrivateGPT
- RAGFlow vs Dify
- LightRAG vs Microsoft GraphRAG
- Docling vs Unstructured (OSS)
- LlamaIndex vs Haystack
- Open WebUI vs LibreChat
- Jan vs GPT4All (LocalDocs)
- Flowise vs Langflow
- Onyx vs PipesHub
- Khoj vs Open Notebook
Flagship setup recipes
Last updated 2026-11-14