Hi, great project as I wanted to give voxtral a try and see if this combo can finally replace whisper.cpp for me.
I know this project is meant for a LIVE / realtime STT, but even for using it to just use the /v1/ for other stuff stuff like Open WebUI, it is really great and thanks to voxtral support it definitely has its quality edge over whisper.cpp
However, currenlty the harcoded 120s timout for the server on the API is limiting. For speaking propmpts of course the 120s is way enough, but transcribing longer audiofiles, WLK currently always gets into the 120s timeout.
It would be very great if the timeout could be either a lot longer or a variable (in seconds).
Hi, great project as I wanted to give voxtral a try and see if this combo can finally replace whisper.cpp for me.
I know this project is meant for a LIVE / realtime STT, but even for using it to just use the /v1/ for other stuff stuff like Open WebUI, it is really great and thanks to voxtral support it definitely has its quality edge over whisper.cpp
However, currenlty the harcoded 120s timout for the server on the API is limiting. For speaking propmpts of course the 120s is way enough, but transcribing longer audiofiles, WLK currently always gets into the 120s timeout.
It would be very great if the timeout could be either a lot longer or a variable (in seconds).