Gradio.Server enables custom frontend with Gradio backend on Hugging Face Spaces

AI Impact Summary

Gradio.Server extends FastAPI to host a custom frontend (React, Svelte, or vanilla HTML/JS) while leveraging Gradio's queuing, concurrency control, and ZeroGPU on HF Spaces. The Text Behind Image example runs a BiRefNet background-removal model behind a @spaces.GPU endpoint and exposes an /remove_background API via gradio_client, enabling a purely client-side UI to orchestrate ML work without you building your own GPU scheduler. This approach delivers scalable, multi-user ML apps with bespoke UIs while maintaining Spaces hosting and programmatic access. Migration considerations include using gradio.Server for full custom UIs or using Gradio components when the UI is simpler.