A marketplace for idle API capacity
Idle provider keys earn. Builders pay a third less. One gateway meters both sides.
Illustrated request flow
01
Spender key
One unified key, budget-capped per project.
02
Gateway
Auth, budget check, router picks a live funder key.
03
Funder key
Real provider key serves the request for its model.
04
Settlement
Every request splits 66/75/25 automatically.
The whole model on one bar. Full pricing →
Live supply with provider list prices struck through and our cut beside them.
OpenAI-compatible endpoint. Pick your SDK — or point your coding CLI at it.
from openai import OpenAI
client = OpenAI(
api_key="sk-your-unburn-key",
base_url="https://api.unburn.quantoxtinc.com/v1",
)
resp = client.chat.completions.create(
model="gpt-4o-mini",
messages=[{"role": "user", "content": "hello"}],
)$49.50
per $100 of direct traffic on your key — capped, pausable, encrypted.
Is my provider key safe?
AES-256 encrypted at rest, never returned by any API, decrypted only server-side to register routes. Pause or delete any time — paused keys get zero traffic instantly.
What does it actually cost?
Spenders pay 66% of direct price per model. Of that, 75% goes to the funder whose key served the request and 25% keeps the platform running.
Which models work?
Anything a funder connects: OpenAI, Anthropic, Google, or any OpenAI-compatible endpoint — named canonically so metering prices correctly.
How do payouts work?
Settlement splits every request and pays funders automatically via Stripe when rails are connected, otherwise earnings accrue. Spender wallets debit pay-as-you-go.
One account funds and spends.