I subscribed briefly, then requested a refund. The reason was both for the highly customisable and curated search, but also for Assistant - because I trust Kagi a little more than Google, Anthropic, OpenAI, or Perplexity with my “best pancake batter according to Reddit” queries - even if that’s fundamentally compromising my privacy. Kagi tended to deliver better information in general, so maybe I’d want to use LMs less frequently - I thought.

However, that slight hint of trust was shattered very quickly when they did what every company with a profit does best - sold me a lie:

https://kagifeedback.org/d/10116-kagi-assistant-standalone-subscription

Which means what I’d subscribed to will have eventually gotten worse, likely for the same price, and Kagi had no intention of letting me know before I subscribed, and their intention seemed to be to just make active subscriptions worse, as they were opposed to grandfathering in the thread.

So I decided I’ll pass, I have enough “products” I don’t trust. Even if it’s all in the “we’re ‘contemplating’ stage”, it’s clear what their approach is and how far they’re willing to go with, pardon me, fucking users over.

P.S. I understand the sustainability problem. But I’d have expected them to have that part figured out before actively offering it in a subscription. The way they are going about it is off-putting. So, no thanks.