Audio feature extraction in the browser

Hi!

Could you recommend ways to extract learned audio features in a JS / web environment, client side?

I’ve found:

  • and see options for audio classification, but could I skip the classification head, using those APIs, to obtain the feature vector / embedding?