Ultramic logo
Ultramic
Create
Roadmap
Feedback
Support

    Boards

  • Support

Powered by Canny

Support

posts
Use ANE models for better performance
For whisper.cpp I optimized the CZ model for CoreML. This improved speed on my M2 like 10%, but with significantly less memory and energy used. It has small priority on the desktop, but could be important on the iPhone. Feel free to contact me, I can convert them for you.
0
Add progress indicator
When recording was stopped nothing visible happens, and then suddenly it pastes the text (it focus was not lost meantime). I propose to show some animation on the icon. The best curcullar progress bar.
0
Transcription is lost when no input is focused
I works when some textfield has focus, but if not the text is lost. It could save it to the clipboard and display notification like: „Your transcription is in the clipboard. Cmd+V to paste it anywhere.“ It could also have last (one or multiple) transcriptions available to copy in the menu.
0
Onboarding transcribe demo does not work when model have not been downloaded yet.
Maybe it just need one more message like: „waiting for local model download"
0
Are accesability permission necessary?
I don't want to give such sensitive permission, but I am ok with doing extra step like Cmd+V after transcription end. Current onboarding forces me to give this permission.
0
Add Linux Support
Enable full functionality of the speech-to-text app on major Linux distributions (e.g., Ubuntu, Fedora, Debian)
0
Powered by Canny