Frequently Asked Questions

How does semantic search differ from keyword search?

Traditional keyword search requires exact word matches. Semantic search understands the meaning behind your query, so it can find relevant passages even when different terminology is used. For example, a search about "baptismal record" can surface results discussing "sacramental register" without that exact phrase being used.

What languages can I search in?

In theory you can search in any language. We use a multilingual embedding model that has Latin and most modern language in its training data. However, current testing shows that you will get better results by searching in the language of the text that you are trying to locate.

How are results ranked?

Results are ranked by semantic similarity to your query, or in other words, the content that seems to be talking about the most similar things to what your query contains. We do not use any generative AI to evaluate the search results (RAG), so there is no attempt at "does this answer the question of the query?", just "does this seem similar in content to the query?"