MOSI Intelligence develops unified-token multimodal foundation models for speech understanding and generation, video understanding, and audio-video generation
MOSI Intelligence develops unified-token multimodal foundation models for speech understanding and generation, video understanding, and audio-video generation. Its models and capabilities include MOSS-TTS, MOSS-Audio, and MOVA, offered through MaaS APIs and the consumer AIGC platform Mossland. The company is incubated by the Shanghai Innovation Institute and Fudan's OpenMOSS team.