The Bavard FAQ Service

When a user visits a site with a question, often then will need to go to an FAQs page and manually scan through a potentially large list of frequently asked questions to find the answer to a question similar to their's. Some websites have the improved functionality of allowing the user to search over these FAQs. Usually this search feature performs a simple exact text match, which means the user has to be very careful about their choice of words in order to find a potential match. This limits the usefulness of the search feature.

The Bavard FAQ Service solves these problems. It is a web service for performing semantic search over an FAQ knowledge base. What this means is that all you have to do is provide your list of FAQs to our service, and we will index those for you, and expose a search endpoint for your FAQ list that your applications can call to let your users search your FAQs easily. The search process is powered by machine learning which allows a user to search semantically: so long as their question has a similar meaning to one of the questions in your FAQ list, the correct answer will be returned. The service also comes with out of the box multilingual support for user queries in 16 languages. This means that even if a user asks their question in a language other than the one your FAQs are written in, the correct answer to their question can still be returned. If you use our FAQ service inside of our chatbot assistants, we handle translation of the FAQ answers for you automatically. This is a powerful feature and immediately turns your FAQ list into a fully multi-lingual semantic search engine.

Below is our full API specification. To see it in a standalone view visit here. Or, if you prefer Swagger UI docs, visit here.

All FAQ list operations in the FAQ service (create, read, update, delete, and search), happen under the context of a Bavard workspace. In other words, an FAQ list is created, managed, and searched over within the scope of a single workspace. To reach an endpoint corresponding to an FAQ list operation, a Bavard JWT token must be included in an authorization header, and the JWT must verify the caller is authorized for that workspace.