Cactus Compute releases a 16.9MB speech model for local CPUs

Co-founder Henry Ndubuaku helped build Whistle into Cactus Compute's existing Needle runtime, adding keyword biasing and word timestamps for on-device applications.

By · Published · Updated

A person speaks into a tabletop microphone beside a compact computer, illustrating Cactus’s Whistle speech model running locally on a CPU.

Co-founder Henry Ndubuaku helped build Whistle into Cactus Compute's existing Needle runtime, adding keyword biasing and word timestamps for on-device applications.