Probably by usage similar to OpenAI since I assume your costs are correlated (compute etc).

You could do a hobby plan which is free up to a certain no. of requests/tokens per day or hour so developers can start building without any friction (I think this is important and sort of expected as most beloved dev tools do it). You can minimize your costs by offering this on shared resources so inference time is a little slow and API can go down at times when there is high usage but its free so users wont mind.

And then have pro plans for higher usage / 99.9999 SLA / fast inference etc (maybe a min subscription with a pay if you go over per 100k tokens option...similar to Vercel with bandwidth and serverless functions).

I have a few thousand things to summarize, and am a paying subscriber. What happens to my account if I automate your page?

I'm happy to pay. I think it's too much but I'd probably unhappily pay $0.05 per summary when they're this good. I'd rather pay $0.005 or less, then I'd plug it into KB tools like Obsidian as well as our bookmarking / link logging workflow firm-wide.

Which reminds me, I want to be able to "gift" Kagi search to all employees. I actively don't want them to have to use Google. Is there a way for an automatic company @domain x per seat billing, or should I use gifting?

Finally, what would it take for kagi.com to be part of the Apple One / iCloud+ offering? That needs to happen.

GPT cannot currently do this (for long documents), and you have a commonly requested use case. For now you can charge as much as they do and have a decent business I'm guessing, but future depends on LLMs from OpenAI and other getting into this space.

I would also like to see controllable length output if that's feasible with your model.