The go-record-to-text is a command line client which records audio from the default input audio device and streams the raw audio data into a speech to text engine.
For audio input, this project uses gordonklaus's go wrapper around the C++ portaudio library
For speech to text, this project uses Mozilla's Deepspeech engine with a pre-trained model, more specifically via asicode's golang deepspeech wrapper
To get this working locally, you must follow the instructions for installing Deepspeech here.