The library extracts a list of words with a timestamp and a duration from a given audio file. It uses OpenAIs whisper.cpp as a backend which can be exchanged easily.
You need a clone of the whisper.cpp repository. Build the whisper.cpp
executable which is simply called main
.
git clone https://github.com/max-and-me/meta-words.git
mkdir build
cd build
cmake [OPTIONS] ../meta-words
cmake --build .
Open source/main.cpp
and adapt the paths according to the cloned whisper repository. Build and run the executable.