Ask about something current and Convaa searches the web and reads pages, citing what it used.
Models have a training cutoff. Convaa closes that gap by searching the web and reading specific pages when a question needs current information, then listing the sources it actually used.
You can also paste a URL and ask about it directly, which is usually faster than copying a page into a prompt and hoping the formatting survives.