Underdog released a browser-based game built with its new text-to-speech model called Woof, which measures just 31.7MB and runs locally without internet. The game features a retro tech environment with a typewriter, CRT monitor, and music tuner that all use Woof for speech generation. Players can type text and hear it spoken back, with the model generating different music styles through the tuner interface.
The model contains 31 million parameters and is available for download. Underdog says Woof can run on most devices without a network connection. The demo comes weeks after the company launched Bark, a 167MB speech-to-text model designed for edge device transcription and dictation.
Sigil Wen, Underdog's founder, told TechCrunch he built the app because he wanted to create his own system with AI models running on users' devices rather than using cloud-based notetakers like Granola. "I think we will be able to run most capable models of today on the devices you already own within months. We just want to train the best open models that are American-made, that are open, that are capable and allow people to have private AI," Wen said.
Underdog has also released a 27-billion-parameter model designed to run on laptops and a waitlisted app positioned as a locally hosted alternative to Muse and Instinct.