Ask AI search for your documentation
The answer, and the page it came from
Documentation search returns a list and leaves the reading to the reader. An Ask AI box answers the question and links the page it used, which is what someone scanning for a parameter actually wanted.
Your docs are complete and your search is keyword-based, so the answer exists on page fourteen of a guide nobody opens. Readers give up and ask in your community, or file a ticket, or leave.
- Your docs site, crawled
- Markdown from the repository
- OpenAPI descriptions and changelogs
- Tutorials and how-to guides
What it does
In the docs, not beside them
The inline surface sits in your own layout, so the assistant is part of the page rather than a bubble floating over it.
Cited to the page
Answers name the page and the section, so a reader can go and read the rest of the context.
Versions stay separate
One knowledge base per version keeps an answer about v2 from quoting v1.
Questions people ask
Does it replace search?
It sits next to it. Search is better when someone knows the term they want; an assistant is better when they only know the problem.
Can I keep it inside my own design?
Yes. The inline surface takes your layout and the widget editor controls the rest, or you can drive it entirely through the React components.
How do I keep it current?
Crawl on a schedule, or push from CI when you publish. A changed page updates its document rather than duplicating it.
Connect it to
Put your knowledge to work today
Free for 14 days. No card, and nothing to install on your site.