Been using it for a year or two, but recently I've been making heavy use of LLM results using the ? character. I actually discovered this by accident while searching. Even though I have a ChatGPT subscription, the Kagi interface is faster to use for simple questions and I have web results underneath as a backup if the LLM answer isn't good enough. I agree with another poster that being able to test ultimate would be nice before I shell out more money.
No comments yet.