Response
ReadyChoose a model and send a prompt.
THE ORDER BOOK
Available intelligence.
Real listings from connected providers. Your own capacity appears here too.
Live offers will appear after the backend is connected.
Response time is measured for a complete non-streaming request, not time to first token. Paid listings can be inspected; payment settlement is not enabled.
YOUR CAPACITY
Connect a provider.
Discover the models attached to your API account.
Connected accounts
Keys are encrypted with AES-256-GCM. Deleting a connection removes its stored credential and listings; revoke the key at your provider as well.
SUPPLY THE MARKET
List your capacity.
One offer per provider key and model. A small completion verifies each new listing.
My offers
A REAL REQUEST
Try your route.
Run text inference through an active own-provider or sponsored offer.
ONE ENDPOINT
Connect your tools.
Create a revocable inference key. Clear shows the complete key once.
Keys authorize inference and model discovery only. Current site access restrictions still apply to external clients; the site remains private until you choose to open access.
OpenAI-compatible request
Read the integration guide ↗YOUR RECEIPTS
See every request.
Measured usage from real verification and inference calls.
Unknown or failed upstream calls retain their conservative token reservation, since the provider may have consumed quota. Dollar figures are estimates, not charges or payouts.