KiteSense raises $28M seed and launches WhisperEdge, an on-device speech model for mobile apps

AI · 4 min read

KiteSense announced a $28 million seed round led by Nightingale Capital and launched WhisperEdge, an on-device speech recognition and intent extraction model optimized for low power mobile and wearable hardware. WhisperEdge claims competitive word-error rates with a fraction of cloud latency and includes a compact acoustic model and lightweight NLU for simple command and dictation use cases.

The product ships with an SDK that abstracts microphone preprocessing, wake-word handling, and on-device privacy controls so user speech never has to leave the device. KiteSense is targeting voice-first apps in healthcare, field services, and accessibility, where offline capability and privacy are key.

Funds will go toward expanding language support, optimizing for additional hardware targets including custom NPUs, and growing partnerships with OEMs for preinstallation. KiteSense positions WhisperEdge as a privacy-first alternative to always-on cloud-based speech services.