I've been using OpenSuperWhisper for development on my MacBook, and I was going back and forth between my MacBook and my PC for looking at code. I found that OpenSuperWhisper was able to increase my productivity by so much, mainly because speaking is three times faster than typing, and so I wanted to just create something that worked similarly on Windows. Since the building blocks for an application like this are all in the open, this should already exist, but I made this and it works pretty well.
WhisperKey
WhisperKey is a Python app that runs on faster-whisper which allows you to run speech recognition locally through OpenAI's Whisper model.
faster-whisper uses CTranslate2, an inference engine, to generate Speech-To-Text (STT) output. An inference engine uses a trained neural network to infer output; in this case, it infers the text to output. I made this to run on my NVIDIA GPU via CUDA with an option to run on CPU.