Earn AIU for the AI
you already serve.
If you run GPUs or host open-weight models, the PRVTC provider model lets you publish capacity, quote in AIU, and settle on signed receipts under private-beta terms.
Three steps to your first settled job.
Publish
Register your models, regions, capacity, and price curves with a receipt-signing key.
Quote
Return signed quotes in AIU for incoming jobs. The SDK handles canonicalization and signing.
Settle
Run the job, sign a metering receipt, and AIU would settle to your account under final beta terms.
OpenAI-compatible. Bring your own backend.
Wrap an existing inference server — vLLM, TGI, LiteLLM, or your own — behind the provider SDK.
// register once, then answer quote requestsimport { Provider } from "prvtc-provider";const p = new Provider({ key: env.PRVTC_KEY }); p.onQuote(async (job) => ({model: job.model, resource: "output_tokens", price_aiu: 0.00000067, // per tokenregion: "eu-central", privacy: "audited", valid_ms: 2000,})); p.onRun(async (job, sign) => {const out = await backend.infer(job); // your serverreturn sign({ output_tokens: out.usage.completion });});
You control supply, price, and privacy.
Models & capacity
Declare which models, how many concurrent jobs, and your realtime vs batch lanes.
Price curves
Set AIU rates per resource, with surge and volume curves you control.
Region & tier
Advertise residency, audited, or on-prem status and charge the premium for it.
Settlement & payout
Track AIU settlement and payout eligibility under final terms. Every job carries a signed receipt.
We start where supply is owned and OSS-friendly.
No need to ask closed providers to accept a new asset. PRVTC begins with operators who own their GPUs and serve open models — then earns the network effect.
Built for independent operators.
- Own GPU servers and clusters
- OSS / open-weight model hosts
- Regional & private AI providers
- Any OpenAI-compatible backend
- Internal boxed-AI deployments
Put idle capacity to work.
Join the provider beta and preview quoting in AIU.