Whistle: Speech to Text in 16.9 MB
Whistle is a new, lightweight speech-to-text model designed for edge devices like wearables and microcontrollers. It operates entirely on the CPU using a 16.9 MB file and shared C++ engine architecture.
Why it matters
This innovation enables high-performance voice recognition on low-power hardware without requiring cloud connectivity or heavy dependencies.
Today we release Whistle, a speech recognition model for mobiles, wearables, robots, smart home, automotive and microcontrollers. It is one 16.9 MB file, runs on the CPU with no dependencies, and loads into the same C++ engine as Needle , from the same container and the same quantisation.
Keywords Press the mic and say something.
Whistle does three jobs, all of them on the device:
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in