Source: Notebook tutorial: LangChain vs LlamaIndex

The owner's comparison of LangChain (orchestration and agents) and LlamaIndex (data ingestion and retrieval), with the same RAG built both ways.

Contents
  • What: interactive HTML tutorial in the owner’s public Notebook
  • Author: Loic (coolloic)
  • Date: undated (read 2026-11-14)
  • Original: LangChain vs LlamaIndex

Summary

Frames LangChain as a broad orchestration toolkit and LlamaIndex as a focused data framework; they overlap on RAG. Builds the same document Q&A in both to show LlamaIndex’s opinionated defaults versus LangChain’s explicit, composable stages.

Key claims

  • LangChain’s limits are sprawl: many abstractions, fast-moving APIs, and more manual RAG assembly.
  • LlamaIndex reaches a working RAG answer in about four lines because it decides the pipeline; LangChain exposes each seam, which costs code but gives control and reuse inside agents and graphs.
  • Rule of thumb: if the hard part is the data, start with LlamaIndex; if it is control flow, start with LangChain/LangGraph; combine them by exposing a LlamaIndex retriever as a tool in a LangGraph agent.

Feeds

LangChain

Suggest changes →