Launched this week
OSPA Fuse
Per-customer AI spend limits for OpenAI, Claude and Gemini
2 followers
Per-customer AI spend limits for OpenAI, Claude and Gemini
2 followers
A proxy for your OpenAI, Claude, Gemini, Mistral and xAI keys. Set dollar limits per project, customer and feature. Every request is priced before it leaves, and the one that would cross a limit gets a 402 and never reaches the provider.
Payment Required
Launch Team
Framer AI AgentsDesign and publish professional sites with AI
Promoted
Hey Product Hunt!
I'm Bonin, the maker of OSPA Fuse.
This year a developer's stolen Gemini key ran up $82,314 in two days, and their billing alerts arrived after the damage. Provider limits are getting better, but they work one provider at a time and they don't know who your customers are. Fuse is a proxy that sits in front of your keys:
• Dollar limits per project, customer and feature, daily or monthly
• Every request is priced at its worst case before it leaves. If it would cross a limit, it gets a 402 and never reaches the provider
• Your app holds a Fuse key, not your real key, so a leak is capped at your ceiling
• Shadow mode lets you test limits before enforcing them
• Works with OpenAI, Anthropic, Google, Mistral and xAI. Setup is two environment variables
It's $19 or $49 a month, with 30 days free on monthly plans. Tokens stay on your own provider bill.
One thing I'd love to know: have you ever had one user, agent or leaked key cause a scary bill? What did you do about it?