For Audiogest I build a custom pipeline to generate transcripts with speaker diarization with Whisper. You can find it here: https://replicate.com/thomasmol/whisper-diarization. This public model has over 75k+ runs, and only about 2k are from Audiogest users. So what I did was create an REST API wrapper around this model, that is Spectropic.
Just send a file or file url to https://api.spectropic.ai/v1/transcripts and receive a transcript with diarization within 5 minutes.
More to follow!