Hi, I discovered Botpress today and I'm really impressed. It would work for my use-case, there's just a point I need to clarify: I was thinking about a endpoint on my website server that would be a proxy to querying Botpress searchFiles endpoint, to create a search engine with "intelligent" results. Website indexed => KB I saw the knowledge base is query-able (searchFiles endpoint) and it did returned some results when I tried. I think I can get from the tags that it's a result from an indexed web-page (dsType = web-page) "tags": { "dsId": "ds-default-website", "kbId": "kb-default", "dsType": "web-page", "source": "knowledge-base", "pageUrl": "https://www.website.ca/crematoriums", "lastCrawledAt": "2026-02-03T19:38:55.654+00:00", "initialBotFile": "true" }, As anyone already used it as a "public" search engine on its website (a search bar giving results) ? Would it be efficient ? Any other idea ? Thanks

yeah it can definitely be done, even the question from the user can query the kb, and you can use the sources to send links to the user so they can navigate the site

Does calling the SearchFile endpoint consume AI credits ?

no: ## Practical rule of thumb - Use

searchFiles

when you want passages from indexed files (RAG building block). - Use KB querying when you want Botpress to search configured KBs as knowledge sources (potentially beyond just files). so if you want to use the botpress KB querying, that will use ai creds, using searchFiles does not.

Thanks, very useful