Instead of letting users buy a monthly package, let them purchase a fixed number of search credits instead, which do not decay.
This will basically be exactly like how the Kagi trial searches work for new users.
Except it can be topped off on demand at any time.
The difference would simply be that instead of offering a plan and either monthly/annual payment, you would offer a third option: Purchase a pack of search credits. Maybe even at a slightly higher price than the equivalent of searches if they'd be bought with a monthly package.
You would not have to promise that the searches/AI features/etc always cost the same number of credits, so you will remain able to adjust your pricing if your costs rise.
Why I consider this to be useful - I've been aware of Kagi for a longer while. I've used up all but one of my free trial searches, and I realize that I don't fit into any of the existing plans.
I would wind up paying 5 bucks a month on the cheapest tier, but then possibly only do 1-2 searches because I'm busy with other things. That means I'd pay $2.5 per search which is absolutely ridiculous.
Other times, I may work on a project that requires a very large amount of searches to narrow a problem down or hunt for something on the web, and I'd blow past the plan's included searches.
From your perspective, having users spend money for little usage is great. More profit. You get money for nothing, shareholders are happy.
From my perspective, having no stable, non-decaying value for the money I'm putting in is terrible and discourages me from paying anything.
I don't consider this to be a greedy or unreasonable idea. "It's only $5 a month dude!"
Yeah, so? Why pay for the potential of doing searches instead of just the searches I actually do?
My account costs you nothing while it sits on your database. I only cost you bandwidth, CPU cycles, electricity, etc., when I fire off a search. And that I'm happy to pay for, just not as a package.
