I’ve really been wanting to flirt more with getting my computer compromised, leaking all my personal and professional data, subjecting myself to prompt injection attacks at every turn and generally making my access to information less trustworthy and my interactions less authentic. Thank you for providing a path to all of these outcomes in a single place.

> Sometimes I find myself seeking answers to random questions, like “how heat propagates through space”, and I often don’t want to google it or switch context to a separate AI chat tool.

Kagi solves this by allowing a question mark at the end and directing the query to an agent in that case.

undefined