A 1-billion-parameter motion model trained on NVIDIA CUDA, requiring 8GB of VRAM for real-time operation. It converts webcam or video footage into real-time XYZ skeletal point data and rotation values, transmitting them to Blender, Unity, and UE for retargeting.

Nov 7, 2025 · 2:44 AM UTC

7
50
1
466
0
Replying to @cyanpuppets
And the name is? And it can be used where ?
4
Your users can sync, share, and manage files — now let them edit them too. Integrate Collabora Online to provide a powerful, open-source document editor directly within your platform.

Ever worried about who can really access your files?

collaboraonline.com
2
Replying to @cyanpuppets
Why doesn't it have fingers
Replying to @cyanpuppets
I'll make my own
Replying to @cyanpuppets
Still needs a f'k load of work with just one camera angle. Watch the character hands spaz out when they go behind the dancer's head.
1
2
Replying to @cyanpuppets
Foss pls
Replying to @cyanpuppets
You are trying to advertise this for years, but not even trying to communicate with us. You are China based and I bet trained your model on our videos that you got from us. Because you don’t have even a UI not a webpage to upload the video by ourselves. Chinaisanasshoe
Load more