Video→3D Pose Tool

Turning a normal video into a moving 3D skeleton is becoming a browser product, not just a studio workflow. VideoMo AI says it can take one clip and output a rigged 3D animation from it. (videomo.ai) The company’s site says the beta is “launching soon” and lists a waitlist of 1,060 users as of April 23, 2026. It says uploads can come from a phone, GoPro, or DSLR camera and that exports include FBX and GLB files. (videomo.ai) The basic idea starts with pose estimation, which means software marks body points in each frame the way a coach might trace joints on pause-screen video. Google’s MediaPipe documentation says pose tools can output landmarks from images, decoded video frames, and live streams, including 3D world coordinates. (ai.google.dev) Animation software then maps those tracked joints onto a digital skeleton, so a recorded walk or jump can drive a 3D character. Three.js, the web graphics library VideoMo references, includes animation clips, mixers, keyframe tracks, cameras, and skeleton helpers for previewing and playing that motion in a browser. (threejs.org) VideoMo’s pitch is not only the body motion but the shot itself. Its site says users can “preview in 3D instantly,” inspect every frame, and export animations at 24, 30, 60, or 120 frames per second, with a static T-pose and a rigged base mesh. (videomo.ai) That places it in a crowded AI motion-capture market that already includes browser tools such as RADiCAL. RADiCAL says it converts 2D video into 3D character animation in the cloud and exports FBX files for Unity, Unreal Engine, Blender, and Maya. (radicalmotion.com) The selling point for all of these tools is time and equipment. Traditional motion capture usually needs multiple cameras or sensor suits, while video-based systems try to recover motion from footage that creators already have. (radicalmotion.com) The hard part is accuracy. VideoMo says it captures full body, hands, and face across 77 joints with “sub-frame accuracy,” but the public site does not publish a benchmark against studio mocap or a technical paper explaining how it reconstructs camera paths from footage. (videomo.ai) For animators, game developers, and web creators, the practical question is less whether 2D-to-3D mocap exists than whether one upload can produce clean motion that survives export. VideoMo is now presenting that workflow as a near-term product release rather than a lab demo. (videomo.ai)