@nixCraft The speech-to-text model runs locally on the user's machine according to this article: theverge.com/2025/1/9/24339817