FreeMoCap v2.0.0-alpha.26

FreeMoCap is a free and open-source markerless motion capture system that allows users to record 3D human movement using standard webcams or video files instead of specialised motion capture suits.

Unlike traditional mocap systems that require body markers and high-end camera arrays, FreeMoCap relies on software-based tracking and computer vision techniques. This makes motion capture more affordable and accessible for students, indie creators, and researchers.

Download FreeMoCap v2.0.0-alpha.26 - Software Mirrors

FreeMoCap v2.0.0-alpha.26 for Windows

freemocap_server_2.0.0-alpha.26_windows-x64-cpu.zip | 306.62 MB

freemocap_2.0.0-alpha.26_windows-x64-cpu.exe | 325.29 MB

FreeMoCap v2.0.0-alpha.26 for macOS

freemocap_server_2.0.0-alpha.26_macos-arm64-apple-silicon.zip | 373.37 MB

freemocap_server_2.0.0-alpha.26_linux-x64-cpu.zip | 530.06 MB

freemocap_2.0.0-alpha.26_macos-arm64-apple-silicon.zip | 414.61 MB

freemocap_2.0.0-alpha.26_macos-arm64-apple-silicon.dmg | 429.65 MB

FreeMoCap v2.0.0-alpha.26 for Linux

freemocap_2.0.0-alpha.26_linux-x64-cpu.deb | 411.13 MB

freemocap_2.0.0-alpha.26_linux-x64-cpu.AppImage | 555.07 MB

FreeMoCap v2.0.0-alpha.26 Source Code

FreeMoCap v2.0.0-alpha.26 Source code (zip)

FreeMoCap v2.0.0-alpha.26 Source code (tar.gz)

FreeMoCap v2.0.0-alpha.26 Release Notes:

FreeMoCap v2.0.0-alpha.26

FreeMoCap v2.0.0-alpha.26 brings improvements to markerless tracking accuracy and reliability, greater control over hardware acceleration, and a number of fixes to data output, logging, and processing. This release focuses on addressing issues reported by our community and making FreeMoCap more reliable across different computers and hardware configurations.

What's New

ONNX Execution Provider Selection

Users can now choose which execution provider FreeMoCap uses for RTMPose inference. The new Execution provider dropdown is available in both real-time tracking and post-processing settings. Depending on the available ONNX Runtime installation, options may include:
  • Auto (automatic selection)
  • NVIDIA TensorRT
  • NVIDIA CUDA
  • Apple CoreML
  • Microsoft DirectML
  • CPU
This gives users more control over hardware acceleration and provides an alternative when automatically selected GPU providers are incompatible with their hardware or software configuration. The available options are determined by the installed ONNX Runtime providers. Note that a provider being listed does not guarantee that the underlying hardware and drivers can successfully initialize it.

Improved MediaPipe Face Tracking

MediaPipe face tracking now uses a head-centered image crop, based on existing facial and head landmarks, instead of performing detection across the entire camera image. This is intended to improve face detection when the subject occupies a smaller portion of the image, particularly in full-body motion capture recordings.

Bug Fixes and Improvements

RTMPose Whole-Body Output Ordering

Fixed an issue where RTMPose whole-body keypoints were being incorrectly separated into hand and face outputs. Previously, the exported left-hand, right-hand, and face files could contain keypoints belonging to different anatomical regions. The updated model definitions and dependencies address this mismatch so the exported data corresponds to the intended keypoints.

Reprojection Error Output

Fixed an issue where reprojection error columns existed in the exported motion capture data but contained no values. Reprojection error data is now passed into the reconstructed skeleton output, allowing users to inspect an important measure of multi-camera reconstruction quality.

Model Download and Initialization Reliability

Improved model initialization behavior to address race conditions that could occur when multiple processing workers attempted to download or prepare the same models simultaneously. These changes are particularly relevant to multi-camera processing and first-time model initialization.

Post-Processing Stability

Improved error handling and resource cleanup in post-processing video workers. These changes help ensure that errors during tracking or video processing are reported more reliably and that associated resources are properly released.

Logging and Diagnostics

Several improvements have been made to logging:
  • Improved saving of application and worker-process logs to disk.
  • Fixed behavior where logs from previous application sessions could remain visible after restarting FreeMoCap.
  • Preserved logs when refreshing the application interface within the same session.
  • Added capture of Python server stdout and stderr in Electron, including errors occurring during server startup.
These changes should make troubleshooting crashes, startup failures, and processing errors easier for both users and developers.

Update Notifications

Improved the update notification interface and update-check behavior. FreeMoCap now supports automatic checks on startup, clearer feedback when checking manually, and dismissible update notifications.

What's Changed

Full changelog: v2.0.0-alpha.25...development

Known Issues

Some known issues remain under investigation: Please report any additional issues through our GitHub repository or Discord community.

Alpha Release Notice

Alpha means Busted! FreeMoCap 2.0 is still under active development. While this release addresses a number of reported bugs, users may encounter unexpected behavior, incomplete functionality, or problems that affect recorded and processed data. We recommend keeping copies of important recordings and checking outputs before using them in research or production workflows.

Installation

Download the appropriate installer for your operating system from the [FreeMoCap download page](https://freemocap.org/download.html). Thank you to all our contributors and community members for helping improve FreeMoCap!
Not sure which build to download? See the download guide: https://docs.freemocap.org/freemocap/download

GPU (CUDA) builds — hosted on Cloudflare R2

(Too large for GitHub's 2 GB asset limit.)

Key Features of FreeMoCap

Markerless Motion Capture

FreeMoCap captures full-body motion without requiring wearable markers or suits. It processes recorded video to reconstruct 3D skeletal movement data.

Multi-Camera and Video Support

Users can connect multiple webcams for better depth accuracy or import pre-recorded video footage. This flexibility allows you to scale your setup based on your budget and technical requirements.

Data Export and Analysis

Captured motion data can be exported in formats such as CSV and NumPy arrays. It also supports workflows compatible with 3D animation tools like Blender. Built-in analysis tools help users inspect skeletal tracking results.

Open-Source Ecosystem

FreeMoCap is fully open source, allowing developers to modify, extend, and integrate it into research or production pipelines. The project benefits from active community contributions and ongoing development.

Performance and Usability

FreeMoCap typically runs in a Python environment and may require installation through package managers. While it provides a graphical interface, initial setup, camera calibration, and environment configuration may require technical familiarity.

  • Accuracy: Motion tracking can be surprisingly detailed with proper lighting and calibration, though results vary based on camera quality and positioning.

  • Processing Speed: Most workflows rely on post-processing rather than real-time tracking.

  • Ease of Use: Setup can be complex for beginners, especially those unfamiliar with Python or camera calibration workflows.

Pros and Cons

Pros

  • Completely free and open source

  • No specialised hardware required

  • Flexible export formats for animation and research

  • Strong developer and academic community

Cons

  • Technical setup may be challenging

  • Limited real-time processing

  • Accuracy depends heavily on camera setup and lighting

Final Verdict

FreeMoCap stands out as a capable and accessible open-source motion capture platform. While it may not fully replace high-end professional mocap systems for commercial film or AAA game production, it provides impressive performance considering its zero cost.

If you’re interested in motion capture but don’t want to invest in expensive suits or proprietary systems, FreeMoCap offers a compelling alternative.

Sources

Official documentation: https://docs.freemocap.org

GitHub repository: https://github.com/freemocap/freemocap

FreeMoCap v2.0.0-alpha.26
Free
Download

Software Informations:
Developer:

Operating System:
Windows / macOS / Linux
Date Added:
2026-11-11T20:07:20.488Z
Categories:

Post a Comment/Report Broken Link: