I've been spending the last couple of weeks building a Wispr Flow clone as an open source project. For context, it is a voice dictation app that lets you type faster, by speaking instead of actually typing. I spent the first week building the basic STT capabilities. One of the coolest features that Wispr Flow has is ASR biasing. Wispr Flow calls it its dictionary. I was able to figure out how to implement that for my project and wanted to share how it was done. What is ASR biasing? ASR ...