Move

What is Move AI?

Move AI is a marker-free motion capture platform developed by Move AI Limited; it utilizes computer vision, deep learning, biomechanics, and kinematic models to reconstruct 3D human movements from ordinary videos.

It does not require actors to wear optical marker garments; its main products include the single-camera iPhone app Move One, the enterprise multi-camera local system Genesis, and the Move API for developers to process videos in bulk.

The old classification of databases, which places Move in the category of AI programming tools, is not accurate; its primary purpose is to generate motion data for use in fields such as film and television, gaming, virtual production, sports, and robotics.

Product line and positioning

Main functions

Single-camera Move One

• Mobile recording: Record an actor on an iPhone without the need for markers or special clothing; free users can record videos up to 30 seconds long, while paid users can record up to 60 seconds long.
• File upload: The web platform allows the uploading of existing MP4, MOV, or AVI files, with a maximum size of 1GB per file; it is recommended to use a fixed camera angle, with a resolution of at least Full HD and a frame rate of 60fps.
• Model selection: In s1, each actor consumes 1 point per second, while in s2 each actor consumes 2 points per second; s2 offers better depth, stability, and foot grounding.
• Processing and previewing: After a video is uploaded, the resulting action is generated; it is possible to compare the input video with the output video, and to check the quality of tracking before exporting.
• Reprocessing: Existing materials can be reprocessed after the model is updated; however, whether additional fees will be charged and what the applicable conditions are depend on the instructions provided in the account settings.
• Team account: It allows multiple users to share points and enables centralized billing management; the specific prices and benefits associated with each membership tier need to be checked on the current account or on the sales page.

Genesis enterprise local system

• Multi-camera reconstruction: It uses multiple synchronized cameras to identify key points on the human body, thereby reconstructing the actor’s movements in three-dimensional space; it supports scenarios with 1 to 8 people and various capture ranges.
• Local processing: Genesis runs on Ubuntu and NVIDIA GPUs, allowing for capturing and processing data locally without relying on the cloud; this approach is suitable for those who prioritize data sovereignty and consistent throughput.
• Camera system: At present, Z CAM is the primary option; a wired connection, professional lenses, and timecode synchronization help to ensure consistency across multiple views. It is recommended to use 6 to 12 cameras for such systems.
• Calibration: It includes internal parameters, external parameters related to the checkerboard pattern, as well as real-time feedback for correction; incorrect lens settings, resolution, or calibration will directly reduce the quality of the footprints, scales, and trajectories.
• Actor Profile: It is possible to create profiles for actors that can be used across different projects, and these profiles can be linked to the tracking results in models such as m1 and m2.
• Real-time and post-processing: Genesis Live is used for live feedback and rehearsals, while Genesis Post is used for high-quality solving, filtering, redirection, and final output.
• Enterprise integration: It allows for the connection of Peel Capture cameras and media assets, and it uses Live Link to transmit real-time footage to Unreal Engine.

Action output and redirection

• 3D animation: It is possible to export files in FBX, BVH, GLB, USDZ, and Blender formats, and it provides both the original motion data and the data after redirection.
• Scientific data: C3D and biomechanical motion data are suitable for sports, research, and analysis of human movements, but they cannot replace proven medical measurement systems.
• Development data: JSON can contain bone hierarchies, joint rotations, and keyframes, facilitating integration into custom applications and data pipelines.
• Video diagnosis: Input videos, render previews, camera-view videos, and visualization files to identify errors and display the results.
• Multiple actors option: Each actor can have its own separate file, or a single file containing all trajectories can be generated; when sphere tracking is enabled, the trajectories of the spheres are also outputted.
• Downstream software: The documentation provides workflows for Blender, Unreal Engine, Maya, Cinema 4D, Unity, NVIDIA Omniverse, MetaHuman, and iClone, among others.

Input and output formats

Move One usage process

1. Install Move AI 3D Motion on the iPhone and create a new Move account; older App Store accounts may need to be logged in through the Legacy portal.
2. Ensure that the actor has given permission for filming and cloud-based processing, remove any other people from the background, and keep the entire body within the frame at all times.
3. Shoot using a fixed landscape camera setup and sufficient lighting; prefer footage in Full HD at 60fps or higher to avoid obstructions, motion blur, and camera movement.
4. Submit videos within the app, or upload MP4, MOV, AVI files to the web platform; then select the s1 or s2 model and confirm the points that will be used.
5. Wait for the processing to be completed, then in the Gallery compare the input, output, and action previews to check for issues such as foot slippage, overlapping limbs, frame loss, and direction problems.
6. Depending on the downstream software, select FBX, USD, BLEND, or motion data; after downloading, carry out bone mapping, redirection, and cleanup.
7. Test the animation in Blender, Unreal Engine, Maya, Unity, or other target tools, and retain the original video and processing parameters for easy rework.

Genesis workflow

1. Based on the capture area, number of people, real-time requirements, and existing equipment, Move AI plans for cameras, lenses, switches, GPU workstations, and software licenses.
2. Install Ubuntu 24.04, specify the NVIDIA drivers, as well as the Genesis application and local processing engine; import the license files that are tied to the hardware and the system.
3. In a wired network, cameras, lenses, resolution, frame rate, and timecode can be configured, while recording and file naming are managed uniformly through Peel Capture.
4. After completing the calibration of the camera’s internal parameters and those of the chessboard, check coverage, focal length, synchronization, lighting, and the capture area before recording the actor’s profile.
5. Record the action using Take, import the camera footage into Genesis Post, and select the number of actors, the model, filters, the rig, as well as the desired output format.
6. Check the tracking and preview, correct any calibration or shooting issues, process the data again, and then deliver the final data to animation, real-time engine, or analysis systems.
7. Regularly install Genesis updates, perform backup calibration, and update actor profiles; also verify the compatibility of existing projects before upgrading the model or Ubuntu.

Price and points for Move One

Points from monthly packages are not carried over and are reset on the anniversary of the subscription; additional points purchased can only be used while the subscription is valid.

The U.S. App Store still shows the old pricing for in-app purchases, such as $14.99 per month and an annual price, which differs from the current web-based subscription cost of $18 plus VAT. New subscriptions are now managed through the website, and any differences due to region, channel, or taxes shall be based on the actual charges applied.

Cancellation and refund

• Web subscriptions can be canceled in the Billing and plans section of the account; once canceled, they remain valid until the end of the current billing cycle, after which renewal stops.
• Old App Store subscriptions need to be canceled in the subscription settings of the Apple device; billing and refunds are subject to the procedures of that respective store.
• The public cancellation policy does not guarantee a proportional refund, and unused monthly credits also expire; it is necessary to check the rules regarding refunds, any excess amounts, and taxes before making a purchase.

Move API pricing

The final unit price is equal to the base rate multiplied by the frame rate, resolution, and multi-camera factor; the duration of the video is charged in the nearest whole second, with a minimum charge of $0.10 per job.

APIs, SDKs, and automation

• The Move API uses GraphQL; the API Key is included in the Authorization request header, and the server enables management of files, Takes, Jobs, cameras, rigs, volumes, and outputs.
• Webhooks send notifications when a task is completed or its status changes, thereby avoiding the need for continuous polling; the receiving end should verify the signature, implement idempotent processing, and prevent replay attacks.
• The official Python SDK simplifies the process of uploading files, creating jobs, and retrieving results. Swift is still listed in the documentation as a feature that will be supported in the future, but there are already public Swift projects for single-camera use available on GitHub.
• The official GitHub organization provides a Python SDK, JavaScript demonstrations, Swift projects, Blender plugins, and recipes; each repository has a different licensing agreement.
• Some examples are licensed under MIT, while the Blender redirection plugin is licensed under GPL-3.0; the fact that open-source SDKs or plugins are available does not mean that the Move AI models, platform, or Genesis software are also open source.
• API Keys should be stored only in server-side environment variables or key management systems; it is necessary to rotate them regularly, monitor their usage, and restrict the materials to which access is allowed as well as the callback targets.

Privacy, Security, and Portrait Authorization

• The privacy policy was updated in August 2025 and applies to Move One, Multi-Cam, cloud and local desktop applications, Move Live, as well as the API.
• The platform may handle contact information, video metadata, the original video, motion data, and animated outputs; motion data can include identifying factors such as height, body proportions, mass distribution, rotation, and speed.
• The policy states that no attempts will be made to identify individuals through videos, movements, or other outputs; instead, data analysis along with AI and machine learning technologies will be used to improve applications, services, marketing efforts, and the overall user experience.
• Demonstrating the product’s capabilities is also listed as one of the possible uses, such as handling video, animations, and outputs; commercial filmmakers should ensure that the actors have given their consent, that there is a contractual authorization, and that it is allowed to display the content publicly.
• The sub-processor page lists AWS, Mixpanel, PostHog, Auth0, Stripe, Datadog, Mailgun, and HubSpot; appropriate safeguards are in place for the processing of data across borders.
• There is no fixed number of days for data retention; it is determined based on the purpose, sensitivity, risks, and legal requirements. The timing of deleting cloud-based data should be specified in the project or corporate contract.
• Genesis’s local processing reduces the need to transfer data to the cloud, but factors such as installation, licensing, updates, telemetry, support, and external integrations still require verification based on the deployment configuration.
• The organization responsible for filming must obtain the appropriate authorization from the actors, the guardians of minors, and the venue, and it must establish controls for accessing, encrypting, storing, and deleting biometric data.

Advantages and limitations

• Actions can be extracted from ordinary videos without the need for special markup, thereby reducing the costs associated with setting up scenes for outdoor shots, sports events, and rapid rehearsals.
• Move One, API, and Genesis cover individuals, platform integrations, and enterprise studios; however, the prices, processing locations, and licensing rules for these three product lines vary.
• Various animation, scientific, and JSON outputs facilitate integration into common production pipelines; however, bone mapping, coordinates, scaling, and animation cleanup still need to be handled.
• A single-camera setup is convenient, but issues such as depth occlusion, rapid rotation, objects that move out of frame, loose clothing, people in the background, and motion blur can reduce accuracy.
• Genesis enables unlimited local processing and real-time preview, but it requires multiple professional cameras, GPUs, a network, calibration, licenses, and skilled operators.
• The API is charged based on the number of seconds of video, the camera used, the resolution, and the frame rate; for projects that involve high-frame-rate 4K multi-camera footage, it is necessary to estimate the cost ahead of processing.
• The prices on the Move One website differ from those for in-app purchases in the old App Store; subscriptions and additional points are subject to the constraints of the available packages, so it is necessary to check each item carefully before making a purchase.

Summary

Move AI is suitable for creators, studios, and development platforms that need to convert real human performances into 3D skeletal movements, rather than being a tool for general AI programming or text generation.

Individuals can use Move One to verify the quality of their recordings, while developers can calculate the costs using the API formulas; companies, on the other hand, should include Genesis’s hardware, calibration services, local data management solutions, and specialized processes in their procurement decisions.