Stellon Labs
Stellon Labs builds extremely small AI models so voice and other ML features can run directly on low-power edge devices. They train frontier “tiny” models optimized for CPU-only inference and minimal memory footprints to avoid cloud dependency. Their first release, Kitten TTS, is a state-of-the-art text-to-speech model under 25 MB designed for real-time on-device playback. Stellon Labs operates as a B2B provider targeting device makers, embedded-software developers, and companies needing on-device ML. The company addresses demand for privacy-preserving, low-latency AI at the edge across consumer electronics and industrial IoT markets.
Stellon Labs
Stellon Labs builds extremely small AI models so voice and other ML features can run directly on low-power edge devices. They train frontier “tiny” models optimized for CPU-only inference and minimal memory footprints to avoid cloud dependency. Their first release, Kitten TTS, is a state-of-the-art text-to-speech model under 25 MB designed for real-time on-device playback. Stellon Labs operates as a B2B provider targeting device makers, embedded-software developers, and companies needing on-device ML. The company addresses demand for privacy-preserving, low-latency AI at the edge across consumer electronics and industrial IoT markets.