Mocai vs XR Animator

Mocai reconstructs a subject’s body, mesh, and motion from one video as markerless motion capture. The maker says footage from any single camera, including a phone, works without markers, suits, or calibration. Each performer has 70 tracked keypoints, and skeleton exports use a 127-joint rig. Capture options include finger articulation, multiple performers, root motion, floor grounding, occlusion repair, adjustable smoothing, and synchronized footage review. Exports include FBX, GLB, BVH, USD, and raw keypoints, with optional retargeting to Mixamo, UE5 Manny, and MetaHuman rigs. Plugins for Blender, Unity, Maya, and Unreal Engine support clip uploads and take imports. A public API supports uploads and webhooks; an MCP server can manage takes and exports. The free plan includes one minute of footage per month, clips up to 15 seconds, and single-person capture. Basic costs $10 a month for 10 minutes monthly and clips up to 30 seconds. The monthly footage allowance is shared between portal and API use, and uploads pause when it is reached.

XR Animator is a free tool for driving MMD or VRM 3D avatars from pose input. It uses MediaPipe and TensorFlow.js to detect poses from webcam video, and supports face, full-body, and combined face, body, and hand tracking. Full-body tracking can use a webcam or an image or video file. Face tracking supports 52 ARKit-compatible blendshapes. Users can record motion capture, export VMD, BVH, and glTF motion, and load VMD, FBX, and BVH motion files; FBX or BVH motion can also be exported to VMD. The scene can include 2D images or video, 3D panoramas, and .x or .glb objects. The web app works in major desktop and smartphone browsers, while Electron apps are available for Windows, Linux, and macOS. Android Chrome supports augmented reality. VMC protocol support and a frameless transparent background for capture apps are limited to Electron mode. The project notes that smartphones with limited processing power may be better suited to face tracking alone.