I'm Linyi, welcome to my self-hosted whisper.cpp app.

I used Blazor Web App to develop this web app. The current language model is the base model provided by whisper.cpp, as my server does not have a GPU. How does it work? Users will upload files to the web server. The file will be saved at a temperary location. The web server will start a process calling ffmpeg to convert the file to mav, then it will call whisper.cpp cli to process the file. Once the file is processed, it will initiate a download automatically. 🖖

Future work:

• Add GPU support.
• Change to large model.