Make Notion search great again: semantic search

In this series, we’re looking into the implementation of a vector index built from the contents of our company Notion pages that allow us not only to search for relevant information but also to enable a language model to directly answer our questions with Notion as its knowledge base. In this article, we will see how to use vector embeddings to search and how to improve its performance.

Numbers, vectors, and charts are real data unless stated otherwise

Last time …

ai article contents implementation index information knowledge knowledge base language language model notion openai questions search semantic semanticsearch series vector vectordatabase