FreeMoCap
freemocap.org
- Free plan
- Open source
- Windows
- macOS
- Linux
- API
- Self-hosted
What FreeMoCap is
FreeMoCap is research-grade software for capturing full-body movement as 3D data without markers. It can use two or more standard webcams, GoPros, or smartphones, with a printed ChArUco board to synchronize camera feeds and establish the coordinate system. You can also record video offline on mobile devices and import it for later processing. Results can be exported in formats including CSV, FBX, and .blend files; a Blender add-on loads processed recordings as a rigged skeleton. Processing happens locally, and the homepage says video data stays on your machine. A backend server download supports headless machines, remote capture rigs, and API use. FreeMoCap is open source under AGPL-3.0, and its current release is alpha. Optional anonymous usage pings may include IP address, timestamp, and camera information; you can disable them in the GUI. The project directs questions to Discord and bug reports to GitHub.
Who it suits
It suits researchers, animators, athletes, and creators who need markerless full-body motion capture. Its offline import, Blender add-on, and headless server option may also suit mobile capture workflows and remote rigs.
Good to know, and what to watch
- Works with webcams, GoPros, or smartphones.
- Imports offline mobile recordings for processing.
- Exports CSV, FBX, and .blend files.
- Processes video locally.
- Free and open source under AGPL-3.0.
- Current release is alpha.
- RTMPose currently lacks working Blender export.
- Optional usage pings can include IP and camera information.
Plans
Every plan and priceFree
FreeMoCap
Free plan
Getting started
- Open the FreeMoCap website
- Download the software for Windows, macOS, Linux, or a backend server
- Set up two or more webcams, GoPros, or smartphones
- Print a ChArUco board and calibrate the camera feeds
- Record video or import offline mobile recordings
- Process the capture and export CSV, FBX, or .blend files
Limits to check
The current release is alpha. RTMPose currently produces 3D data without working Blender export.
The verdict
FreeMoCap offers a broad capture workflow, local processing, and several export options at no cost. Its alpha status and the RTMPose Blender-export limitation are worth considering before adopting it for a project.
Questions people ask
Is FreeMoCap free?
Yes. The Free plan is free and has no subscriptions, licenses, or hidden fees.
Which platforms does it support?
FreeMoCap supports Windows, macOS, Linux, API use, and self-hosted deployments.
Does video leave my computer?
Processing happens locally, and the homepage says video data never leaves the user’s machine.
Can it work with Blender?
A Blender add-on loads processed recordings into Blender as a rigged skeleton. RTMPose currently does not have working Blender export.
What cameras can I use?
It supports two or more standard webcams, GoPros, or smartphones, with ChArUco-board calibration.
What the maker publishes
- API
- A local REST API exposes camera detection, recording, playback, calibration, mocap processing, realtime pipelines and Blender operations. source
- Architecture
- FreeMoCap uses a React and TypeScript Electron frontend communicating with a Python backend over REST and WebSocket. source
- Audience
- The project says it is built for researchers, animators, athletes, and creators worldwide. source
- Blender integration
- The documentation lists a FreeMoCap Blender add-on that loads processed recordings into Blender as a rigged skeleton. source
- Blender integration
- The application can install its Blender add-on, detect Blender, export recordings to .blend files and open them in Blender. source
- Calibration
- Calibration uses a printed ChArUco board to synchronize feeds and define the 3D coordinate system. source
- Calibration
- A printed ChArUco board is waved through the camera views to synchronize feeds and establish the 3D coordinate system. source
- Camera inputs
- The official workflow supports two or more standard webcams, GoPros, smartphones used as cameras, and imported mobile video. source
- Camera support
- It supports two or more standard webcams, GoPros, or smartphones as cameras. source
- Company facts
- LinkedIn lists FreeMoCap Foundation as a nonprofit software-development organization founded in 2021, headquartered in Boston, Massachusetts, with 2–10 employees. source
- Company legal status
- The project is managed by the FreeMoCap Foundation, a Massachusetts 501(c)(3) nonprofit corporation. source
- Current release
- The latest listed release is v2.0.0-alpha.25, published 24 September 2026, with CPU startup, Linux server, Blender export and hand-detection fixes. source
- Data protection
- The privacy policy says collected data is not sold or distributed to third parties except as required by law and is protected with industry-standard security measures. source
- Documentation
- Documentation covers installation, camera setup, architecture, API endpoints, development builds and troubleshooting, but the API pages are marked AI-generated and may contain inaccuracies. source
- Export formats
- Results include CSV, FBX, and .blend files, among other formats. source
- Funding model
- The project says it is funded through external grants and donations, while labor-intensive setup, support, training and custom work are paid services. source
- Headless use
- A backend server download is provided for headless machines, remote capture rigs, and API use. source
- Independent listing
- AlternativeTo lists FreeMoCap as free, open source under AGPL-3.0, supporting Linux, Windows, Mac and Python, with no comments or reviews on that listing. source
- Licence
- The project is licensed under GNU Affero General Public License version 3 (AGPL-3.0). source
- License
- The download page identifies FreeMoCap as open source licensed under AGPL-3.0. source
- Limit: alpha quality
- Release notes warn that the alpha may have rough edges, missing functionality and data-affecting bugs. source
- Limit: calibration burden
- A peer-reviewed gait study reports that ChArUco calibration takes operator training, about 1–2 minutes of board movement and 5–10 minutes of processing, with repeats if convergence fails. source
- Limit: maturity and adoption
- An independent technical review describes the project as alpha software and unsuitable for users needing a stable production pipeline or permissive closed-source distribution. source
- Limit: scale accuracy
- The current release notes warn that reconstructed skeleton scale may be wrong by roughly 10 percent. source
Checked against the maker's pages on 15 Oct 2026.