Hugging Face Inference
Biyan supports Hugging Face as a remote inference provider, using an approved Inference Provider/Router API or a dedicated Inference Endpoint.
- Create a Hugging Face token with only the permissions required for inference.
- Confirm billing and access for the remote service you plan to use.
- Open Settings → Model Providers → Hugging Face.
- Enter the token and, for a dedicated endpoint, its exact URL.
- Add or select a remotely served model ID and test it.
⚠️
This integration does not browse or download model weights. Biyan sends Hugging Face requests to the configured official endpoint and does not rewrite them to an application-owned mirror.
Availability, Provider selection, context limits, data handling, and charges are controlled by Hugging Face and the selected inference provider.