MOSI Intelligence
Shanghai multimodal AI lab building real-time speech, vision and interaction models
About
MOSI Intelligence (Shanghai Mosi Intelligence Technology Co., Ltd., 模思智能) is a Shanghai-based large-model company focused on context-aware, real-time interactive intelligence. It was incubated from Fudan University's OpenMOSS team and Shanghai Chuangzhi College, with Professor Qiu Xipeng as chief scientist. Its model family spans speech generation (MOSS-TTS, MOSS-TTSD), transcription and diarization, visual understanding (MOSS-VL) and synchronized video-and-audio generation (MOVA), offered through the Moss Open Platform API. On October 12, 2026 it announced a Pre-A round of over 1 billion yuan led by Huakong Fund, following an angel round in April 2026, and previewed its real-time interaction system MOSS Live.
- Website
- mosi.cn/en
- Added via
- web
- Ownership
- unclaimed
Where to go
Products
- Moss Open Platform — API console for MOSI's MOSS speech, vision and video models · Api
- MOSS-Transcribe-Diarize — End-to-end model for transcription, speaker diarization and timestamps · Other
- MOSS-TTS — Long-form multilingual speech generation model with zero-shot voice cloning · Other
- MOSS-TTSD — Multi-speaker dialogue speech generation from character scripts · Other
- MOSS-VL — Open-source visual understanding model for images and long video · Open Source
- MOVA (MOSS-Video-and-Audio) — Foundation model for synchronized video and audio generation · Other