Custom Denoising Model

Integrate your own trained audio denoising model.

Process Audio with Your Model
Upload an audio file and provide the endpoint for your deployed denoising model.

Click to upload or drag and drop

audio/*

The URL where your Python inference server is hosted (e.g., on Cloud Run).

How It Works
  1. First, train your own audio denoising model using the Python scripts and instructions you provided.
  2. Deploy your trained model as a web service (e.g., to Google Cloud Run). This will give you an endpoint URL.
  3. Paste that URL into the "Custom Model Endpoint URL" field above.
  4. Upload an audio file. When you click "Process Audio", the app simulates uploading the file to a cloud storage bucket.
  5. It then calls a Genkit flow, which in turn calls your deployed Python model's API with the storage location of the audio.
  6. Your model processes the audio and saves the result back to cloud storage, returning the new location.
  7. The result is then displayed here for you to play and download.