Is it possible to get the words from this library or transcription?

[orionflame]

Hi,

Basically this only detects voice right? So there are no transcription capabilities? I want to use it to edit the audio to remove retakes. So VAD helps for the timing, but I also need to know exactly what's being said.

I used whisper and stable-ts but the timings are not so accurate and they suggested it's not suitable for editing.

If I use whisper to feed the entire transcription into Chat GPT, it's able to remove the retakes from the text but I need the accurate timing info so I can overlay the last retakes onto first ones to replace them without changing audio length.

Any ideas?

Thanks a lot in advance.