Vosk is an open-source speech recognition library that provides offline, real-time speech-to-text conversion (STT). Unlike some cloud-based services, Vosk operates locally on your machine, offering a secure and flexible solution for voice-driven applications. While it lacks some advanced features like automatic punctuation, it is highly customizable and can be integrated into a wide range of projects.
How to change models
Check out this video for instructions on how to change Vosk stt models.