VoiceStreamAI
VoiceStreamAI copied to clipboard
Near-Realtime audio transcription using self-hosted Whisper and WebSocket in Python/JS
When another strategy is selected, the old strategy's panel should hide, but it actually does not. Incorrect interation: https://github.com/alesaccoia/VoiceStreamAI/assets/39730824/5c98a25c-5210-4ded-a99e-28712d2f0825 Correct interation: https://github.com/alesaccoia/VoiceStreamAI/assets/39730824/d5a8f8f9-2f9e-4842-9334-bf0caad65a6d
https://github.com/alesaccoia/VoiceStreamAI/assets/39730824/acd2bb0d-3714-4509-9339-13792bf39902
When I want to debug, I need a command option to pass the log level to obtain logs from the WebSocket library. I think we should add an option such...
Loading the OpenAI Whisper model into the GPU firstly instead of RAM using CPU. Currently, our app only loads the OpenAI Whisper model into RAM using the CPU, instead of...
The `ScriptProcessorNode` has been marked as a deprecated feature and needs to be rewritten using `AudioWorkletProcessor`. [demo](https://googlechromelabs.github.io/web-audio-samples/audio-worklet/migration/worklet-recorder/) [code](https://github.com/GoogleChromeLabs/web-audio-samples/tree/3847f84bb9dd5f998c6efb61ccffe9e9f15df119/src/audio-worklet/migration/worklet-recorder)
Is VoiceStreamAI compatible with Azure OpenAI Whisper?
Django?
I'm interested in incorporating this into a Django project. Has anyone adapted this to work with Django and Channels? Thanks!
are there any DEBUG or logging settings? the Docker builds with no problem and starts up. then we start the client. it is not clear it is doing anything but...