- Gradio - web UI for voice-to-action demo

An open-source Python framework for building interactive machine learning interfaces (including voice-to-action agents) with zero frontend experience required.

Gradio transforms Python scripts into interactive web applications for voice-controlled AI. By wrapping models like OpenAI's Whisper in the `gr.Audio` component, developers create hands-free interfaces that process 16kHz microphone input in real time. The framework pipes transcribed text into LLMs to trigger specific actions (like API calls or hardware commands) while managing the frontend state automatically: you focus on the logic while Gradio generates a shareable public link. It is the go-to tool for prototyping voice-to-action agents at companies like Hugging Face and AWS.

https://www.gradio.app