Add Ability to Attached Audio Files for Whisper and AI Model Processing

[slashdotblake]

Koboldcpp is nice in that it supports image attachments. If it could progress to supporting audio file attachments such as MP3 and WAV files that would make this tool even more useful. Really want the ability to transcribe audio to text and to ask AI questions regarding the content of the audio file.

[LostRuins]

It does have this ability!
https://github.com/LostRuins/koboldcpp/wiki#what-is-whisper

Grab a whisper transcription model here https://huggingface.co/koboldcpp/whisper/tree/main and load it in.

Enable voice control in KoboldAI Lite UI

You can also transcribe a wav file.

[slashdotblake]

Thank you for the response. Going to try this now. Is it possible to add a button from the main screen similar to IMG for easy access?

[LostRuins]

Yes if you enable push to talk, your submit button will turn into a mic which you can use to talk to the ai.