Two of us spent weeks building an AI app builder with no credits to sell. Smart or stupid?
Hey PH 👋 We're Baptiste & Samuel, and next week we're launching Capacity Desktop — a free Mac app that turns plain English into real apps, built and stored on your machine.
The bet everyone tells us is wrong: move away from credits system. You plug in your own Claude / GPT / Open Router key and pay your provider at cost. Our take is that you feel the credits system is unfair, you never know what a failed fix loop just cost you, and you feel the tool profits from your retries.
The obvious counter: credits are simpler. One checkout, no API keys, beginners never think about tokens. Bring-your-own-key is real friction, and we know it.
So before we launch tomorrow, genuinely curious:
Would bringing your own API key stop you from trying an AI app builder?
If you use Lovable / Bolt / v0 — does credit anxiety actually change how you build?
Does "the code lives on your Mac, in your git repo" matter enough to make you switch — or is that just something devs say?
Brutal answers welcome — they decide what we fix first. We launch tomorrow morning; hit follow on the page if you want the ping. We'll be in the comments all day 🙏
Replies
Very cool take, excited to see the launch!
I'd rather connect my api key yeah, credits (on lovable for eg. ) have always been quite obscure to me!
You can see buzz (jack dorsey's new agentic slack startup) is doing something similar in their onboarding- thinks it's super smart!
as for the repo, as long as it's backed up on github i'm good
hope it helps! good luck for the rest!
DO credits really change how people build? o
Capacity
@casper_voigt It does change how people feel which impacts the perception they have of the product. Also, considering AI subscriptions, AI API keys, and credit systems, it seems that some app builders rolled out a third business model that is way more expensive. Relying on API keys only seems fairer to us.
You already named the real tradeoff, so I will not pretend otherwise. BYOK friction is a conversion problem, not a philosophy problem. People do not object to paying their own provider, they object to going to platform.openai.com, generating a key, and pasting it into an app they just met.
Being upfront that I have a stake here, I have been building the boring half of that as a service (monet.gg). Instead of a paste-your-key field, the user hits a hosted consent page and clicks Connect on their OpenAI account, like a normal OAuth flow. Your app receives a bearer token and never the key itself, and you call one OpenAI-compatible endpoint. The economics stay exactly what you designed, the setup step stops looking like homework.
On your third question, my read is that "the code lives on your Mac" matters enormously to people who already have an opinion about it and roughly zero to everyone else. I would not lead with it.
Free through beta if you want to poke holes in it: demo.monet.gg