LlamaIndex
RAG-focused framework — primitives for data ingestion, chunking, indexing and querying over your own data, from PDFs to SQL.
Jerry Liu, 2022 (started as "GPT Index"). Has stronger opinions about what RAG pipelines should look like than LangChain: Document → NodeParser → Index → Retriever → QueryEngine. Supports vector, summary, knowledge-graph and compositional indexes.
Good fit when the user's input is queries against a document corpus (internal wiki, legal, support). LangChain is more general; LlamaIndex is the RAG specialist.