Speech Recognition Without the Internet Now Fits on a $10 Chip (Whistle)

Why we like it

Better Stack fits Whistle, a 16.9 MB speech-to-text model, onto a $10 ESP32-S3, where it runs with no Wi-Fi and no cloud. They walk through what they cut to make it fit and how they brought processing time from 24 seconds to about 10, so it's not instant. The full project is on GitHub if you want to build one.