I built Ceptile because I wanted a search tool that could combine the convenience of an AI answer with the ability to inspect the actual web pages behind it.
A Ceptile search retrieves pages from the web, ranks the retrieved sources, and uses them to produce an answer with links back to the sources. You can open the sources and continue researching from there.
I've also been building some things around the search workflow, including Notes for keeping research, translation tools, and ways to work with information found during a search. Just select a text in answer or web results, and it will show a toolbar to save it to notes, translate, etc.
The project is still early. A lot of the work so far has been figuring out the search pipeline itself: retrieving useful pages, dealing with noisy results, deciding which sources should be used, and keeping the connection between an answer and its sources.
I'm particularly interested in where the current system breaks down. For example, if you try it with a query that you already know well, I'd be interested in whether the retrieved sources are useful and whether the citations make it easy to verify the answer.
Watch 2 minutes demo on Youtube: [https://www.youtube.com/watch?v=OR4WfsMKEiE](https://www.youtube.com/watch?v=OR4WfsMKEiE)
Comments URL: [https://news.ycombinator.com/item?id=49812270](https://news.ycombinator.com/item?id=49812270)
Points: 1