LlamaIndex
A Python framework for retrieval only. It connects a model to your own documents, databases or APIs, and does not try to chain everything.
Tested on Claude Sonnet · 4.5 ·
LIBRARY · Free, open source
LangChain covers chaining broadly. LlamaIndex is narrower and mostly about getting your own data in front of a model properly, with indexing strategies, chunking options and retrieval evaluation tools built for that one job. If retrieval-augmented generation is the actual task, the smaller scope usually means less to learn.
Take the evaluation seriously. The framework ships tools for scoring retrieval quality, and running them on a sample of real questions exposes a bad chunking strategy long before a user does.
Notes from the author
Run the built-in retrieval evaluation on real questions before you ship. A chunking strategy that looks fine on one example query can fail silently on the rest.
For information only, not professional advice. You use it at your own risk and we are not liable for any loss or damage. See the disclaimer.
Tagged