MovementModeler · iOS
Video in. Motion data out.
MovementModeler turns any video into clean stick-figure motion and exportable BODY-25 JSON with embedded metadata, all on the phone.
What it does
Stick-figure visualization
Raw footage becomes a clear motion clip, drawn as sticks, capsules, or a sprite rig, so posture, timing, and form are easy to see. Smoothing, track IDs, and confidence thresholds are adjustable in the app.
BODY-25 JSON export
Structured pose data per frame and per person, with coordinates and confidence for every joint, ready for computer vision and machine learning pipelines with no server setup.
MP4 export
Clean stick-figure, capsule, or sprite renders for reviews, teaching materials, or sharing.
On-device and private
No uploads, no accounts, no cloud. Multi-person tracking, smoothing, and gap filling all run on the iPhone.
Metadata, before export
Each capture can be tagged in the app before it leaves the phone, so the JSON arrives downstream ready to sort, filter, and process.
- Activity. A built-in vocabulary (gait, sit to stand, balance, dance, upper limb, stairs, and more) or a custom label.
- Camera angle. Frontal, sagittal, or other, so downstream scripts can handle perspective-specific analysis.
- Subject. In multi-person scenes, which tracked person is the primary subject. The index is written into the export.
{
"fps": 15.0,
"lmm": {
"activity": "sit_to_stand",
"subject_index": 0,
"camera_angle": "sagittal"
},
"frames": [ ... ]
}
From video to labeled data
Record or import a clip, extract the motion, check it visually, tag it, and export JSON, MP4, or both. Everything happens on the device.
Who it is for
Coaches reviewing form with stick-figure clips and sharing short MP4 summaries. Researchers who want BODY-25 JSON with metadata for their own workflows, with the data kept local. Creators and educators turning choreography and demonstrations into clear visualizations. There is a separate page for instructors.
It is also the capture tool behind our own research.