AIwithhassan/ai-skin-specialist ? reverse-engineered prompt

Build me a simple Gradio web app for a skin care assistant that can take a patient’s voice description and a skin photo, then turn that into a clear doctor style response.

I want the app to transcribe the spoken input, look at the image, and generate a helpful written answer plus an audio reply that sounds like a doctor speaking back to the user. It should feel like a complete end to end demo, with a clean interface where someone can record or upload audio, upload a skin image, click analyze, and see the transcript, the guidance, and the spoken response.

Please keep it honest that this is for general informational guidance only and not a real diagnosis. Use the model and audio services that fit this setup, and if you need to check current docs online, go ahead. Make the app easy to run locally with environment variables and a clear sample config file.