Familiar builds world models for accurate humans, offering a dubbing model that translates videos and livestreams into 25 languages while preserving the original voice and lip movements in real time. Its research spans joint audio-visual generation (JAM-Flow), real-time human animation (SoulX-LiveAct at 20 frames per second), and fast diffusion video generation. The team comes from Stanford, Yonsei's Visual Intelligence Lab, and HKUST, and the company describes itself as one of the most accurate video translators available and the only one operating in real time.
Familiar offers a unique dubbing model that translates videos and livestreams into 25 languages while preserving the original voice and lip movements in real time. Key features include:
The company also offers a free plan called Familiar Alpha, which includes 24,000 credits per month for up to 4 minutes of video in one target language, and a Casual plan priced at $35 per month, targeting content creators and businesses.