Projects

Have you ever remembered the content of a Bible verse but not where it is found? Or have you tried searching for a verse using a traditional keyword search, only to realize you couldn't remember its exact wording? If so, this project may be able to help.

Unlike a traditional keyword search, this search is based on semantic similarity. Instead of entering only a few words, you can describe the passage you're looking for in a sentence or two. Your query is converted into a vector and compared with the vectors of individual Bible verses, or if you prefer, blocks of five consecutive verses. The results are then ranked by how closely their meaning matches your query.

For the best results, describe the passage you're looking for as accurately as possible. You don't need to remember the exact wording of a verse; recalling its main idea is enough. Additionally, the vector embedding model performs best with English.

The search is powered by PostgreSQL with the pgvector extension, and the vector embeddings are generated using OpenAI's text-embedding-3-small model. Additional information about the project, supported Bible translations, and the source code is available on GitHub.