Move.ai's marker-free capture still needs a camera rig

The shot

Move.ai documents a markerless motion-capture system it says it has developed since 2019, and its current technology page and products page, retrieved 16 September 2026, describe a multi-camera, GPU-based pipeline named Genesis rather than a single-phone method.

What the documents show

The technology page states the method plainly: “Markerless motion capture uses multiple cameras with computer vision technology to understand and capture motion data purely from video,” working in “any environment,” with “no markers or suits,” across capture volumes from 4x4 metres to 20x20 metres. It credits “patented kinematic understanding of human physics and biomechanics,” analysed frame by frame across the camera array — a vendor description of a proprietary method, not an independently audited one. The products page is specific about hardware: it recommends Z-Cam cameras over consumer action cameras, stating that “consumer action cameras lack the synchronisation and professional lens options required for optimal captures,” and states the Genesis software needs a local Nvidia GPU (RTX 30, 40, 50 or A-series) rather than cloud processing, calibrated with a checkerboard method the page says outperforms a human-based alternative. On the quality comparison this site treats cautiously, Move.ai makes the claim itself: its accuracy page states the company “continuously benchmarks the quality of our system against Optical and inertial systems,” running its own rig alongside a Vicon or OptiTrack setup on the same subject, including a named comparative test at VIP Studio conducted by IJAD Dance Company. That is Move's own self-reported benchmark, not a third-party publication.

The workflow

For an independent production, the documented setup is still a multi-camera install, not a single handheld phone: a studio configures cameras and a local GPU workstation, calibrates the volume with a checkerboard, and can mix Move's system with existing Vicon or OptiTrack cameras in a stated “hybrid” mode rather than replacing them outright. Recording happens locally, which Move's own pages frame as a data-sovereignty choice, and Peel Capture integration lets the cameras start and stop from an existing session manager rather than a separate app.

What the tool does not change

Move.ai's own comparison claim is a vendor benchmark against named competitors, not an independent study this site can verify, and readers should treat the optical-quality language accordingly; the cited comparative test names a venue and a dance company but not a published, reviewable methodology. Choosing a capture volume, calibrating correctly, and cleaning the resulting animation for a shot's specific needs still requires a trained operator the documentation does not describe replacing.

• Does the setup described actually require a multi-camera install and local GPU hardware, or does the coverage assume a simpler single-camera method?
• Is a stated quality comparison to marker-based systems the vendor's own benchmark, or has it been reviewed by an independent party?
• What does the hybrid-compatibility claim mean for a studio's existing camera investment, and is it documented as full interoperability or a specific processing option?

Move.ai's own materials describe a capable but still equipment-dependent system, benchmarked by the company itself against the marker-based rigs it aims to substitute for; that self-reported comparison is worth naming as such before a lower-budget production plans around it.