This app helps use podcasts in different languages you like to learn, downloads the episodes, transcribes the episodes with the help of WhisperCpp and displays the transcript to in combination with the audio. So you can listen and reading at the same time ... and learn.
How to run this app:
This app is based on ruby on rails. So you need a funtioning Ruby installation (Version 3.4 would be the best). You also need whispercpp.
Execute the following to init whispercpp
git submodule update --init --depth=1
after that go to vendor/whisper.cpp and execute
mkdir build
cd build
cmake ../
make
After that you should have whispercpp-cli in vendor/whisper.cpp/build/bin/
The easiest way to run this is with the help of Podman or Docker. In this README the examples use podman but you can exchange it seamlessly with docker. To take advantage of this just run
podman compose -f docker-compose.yml up
you need to download a model for the whisperserver - so far base.en.bin is enough for testing Here you can download it: https://huggingface.co/ggerganov/whisper.cpp/tree/main put it into storage/whisper/models