Frequently asked questions
Short answers to common developer questions. The API reference has the full details.
Is it really OpenAI-compatible?
Yes. The chat-completions and models endpoints follow the OpenAI schema, so official OpenAI SDKs work without changes. Set the base URL to the AI HUB endpoint and use a Hub key. No dedicated AI HUB SDK or CLI is required.
Which models are available?
Any model AI HUB has published, listed live in the Models table (add ?type=image, audio, or video to see media models). The API generates text, images, audio and video. See Image & video.
How is usage billed?
Each call is billed at the model's list price for the tokens used and debited from the prepaid USD wallet when the response completes. There is no separate API invoice and no per-model subscription, because API usage shares the wallet and prices of the rest of AI HUB.
What happens when the balance runs out?
Calls return 402 insufficient_quota until a top-up is made on the billing page. Nothing is queued or charged afterwards. After a top-up, the next call goes through.
Can a key's spend be limited?
Yes. An optional monthly spend cap can be set when a key is created. Past the cap, that key returns a 402 while the rest of the account keeps working. This is useful for isolating a project or sharing a bounded key with a teammate.
How is a key rotated or revoked?
On the Developer page. Revocation is immediate, and any app using that key stops working at once. To rotate, create a new key, switch the app over, then revoke the old one.