Whistle An ultra-compact 16.9 MB open speech-to-text model that runs entirely on CPU with no dependencies, transcribing up to 30 seconds of audio in seven languages with word-level timestamps. Built to share the Needle runtime for on-device audio-to-action pipelines across mobiles, wearables, robots, and 17 platforms.
Visit productCosigns Received
Be the first Cosign
Put your name behind Whistle.
Add it to your Products I use list and share why you use it.