README.md

These models are for KoboldCpp's speech recognition. You would load them with --whispermodel or from the "Audio" tab in the GUI launcher.

Then in KoboldAI Lite, you can enable Voice Input in the settings (Push To Talk or Voice Detect)