LangChain

A library for building LLM-based applications.

Combined with Vector database like Pinecone, it provides a sort of “external memory” for LLMs.

The idea is that if we store the embedding of small chunks of texts in a vector database, then we can use the embedding of the user query to retrieve the most relevant text chunks from the database, which can then be fed into the query as a part of the context.

Tutorials