🤖AI-generated documentation☐ curatedAI Generated

This page was drafted by an AI assistant and may contain inaccuracies.

About content generation types

🤖

AI Generated — Page drafted entirely by AI from codebase or prompt instructions.
(e.g., docs generated from codebase analysis)

← this page

✋→🤖

AI Transformatted — Human provided raw material; AI restructured it into a different format.
(e.g., livestream → blog post, meeting notes → docs)

✋

Human Generated — Page written entirely by a human author.
(e.g., hand-written tutorial)

More info about content generation types ↗

Blender Export

FreeMoCap can export processed mocap data directly to Blender, creating a complete 3D scene with skeleton mesh, camera frustums, and lighting.

MediaPipe only (for now)

Blender export currently works with MediaPipe mocap output only — RTMPose 3D data does not yet have a working Blender export. See the Post-Hoc Motion Capture guide.

Prerequisites

• Blender installed on your system (detected automatically or specified manually)
• A processed recording with mocap data (all required .npy files present in the recording's output data directory)
• The freemocap_blender_addon (installed automatically by the backend)

Export Flow

1. Process mocap data — Run the mocap pipeline on a recording to generate .npy files
2. Detect Blender — The backend auto-detects installed Blender via GET /freemocap/blender/detect
3. Export — Click "Export to Blender" or call POST /freemocap/blender/export
4. Open — Optionally auto-open the .blend file in Blender's GUI

What Gets Exported

The Blender addon reads the .npy output files and builds a complete 3D scene:

Element	Source Data
Skeleton mesh	mediapipe_body_3d_xyz.npy — body keypoints connected as a skeleton
Hand meshes	mediapipe_right_hand_3d_xyz.npy, mediapipe_left_hand_3d_xyz.npy
Face mesh	mediapipe_face_3d_xyz.npy
Center of mass	mediapipe_body_total_body_center_of_mass.npy — whole-body COM trajectory
Camera frustums	Calibration data — shows where each camera was in 3D space
Ground plane	Calibration groundplane alignment — Z=0 reference plane
Lighting	Default three-point lighting setup

Technical Details

The export process runs Blender in --background mode:

1. Backend resolves the .blend output path from the recording folder
2. Injects freemocap_blender_addon into Blender's sys.path (no manual addon installation needed)
3. Runs run_blender_export.py via --python
4. The addon reads .npy files, builds the scene, and saves the .blend
5. Backend verifies the .blend file was created

API Endpoints

Endpoint	Purpose
GET /freemocap/blender/detect	Find installed Blender
POST /freemocap/blender/addon/install	Install addon into Blender
POST /freemocap/blender/export	Export recording to .blend
POST /freemocap/blender/open	Open .blend in Blender GUI

Troubleshooting

"Blender not detected" → Install Blender from blender.org, or specify the path manually in the UI settings.

"Export fails — missing .npy files" → Run the mocap processing pipeline first. The export readiness check verifies all required files exist.

"Addon failed to load" → The addon is injected at export time, not installed into Blender's user preferences. Ensure the freemocap_blender_addon package is available in the Python environment.