sammykumar-ComfyUI-SwissArmyKnife/web/docs/VACEDepthAnnotator.md

VACEDepthAnnotator

Run the VACE depth estimator to produce H×W depth maps for each frame, with automatic fallback when checkpoints are missing.

Inputs

• images – IMAGE batch to analyze.
• style / inference_mode / resolution – control checkpoint selection and processing size.
• Optional model_path – override path discovery.

Outputs

• depth_maps – IMAGE tensor containing normalized depth values.

Usage Tips

1. Drop into video pipelines before ControlNet/conditioning nodes.
2. Keep resolution reasonable (512) unless you need extra detail.

Additional Resources

• Full documentation