Inference Endpoints can be used through the UI and programmatically through an API. Here you’ll find the open-API specification for each available route, which you can call directly, or through the Hugging Face Hub python client.
< > Update on GitHub