Connect Hugging Face models in one click

Deployments can start from models in your Hugging Face account. Connect once via OAuth, browse your repos and public models, and submit a serving plan without hand-writing deployment configs.

The planner estimates time-to-ready, serving stack, and cost before you approve. After approval, orchestration provisions capacity, starts the inference server, and publishes your endpoint URL.

Amazon S3 and Google Cloud Storage model sources follow the same connect-and-plan flow. More providers are coming soon.