Video Depth Map Generator
title
subtitle
upload_video_tip
colormap_tip
Extract a stable depth map from any video. Depth is normalized once across the whole clip, so subjects keep the same gray level as the camera moves and nothing flickers frame to frame.
Why use the video depth map generator?
Consistent across the entire clip
Depth is normalized over the full video rather than frame by frame, so there is no flicker or brightness pumping.
Sharp subject separation
Clean silhouettes on people, hands, and thin structures, with background geometry preserved instead of crushed to black.
Ready for control workflows
The default grayscale output (near = white, far = black) drops straight into depth-conditioned video models and ControlNet-style pipelines.
Original resolution and frame rate
The depth video matches your input resolution and frame rate (up to 30 fps), so it lines up with the source in any editor.
Three visualization styles
Keep Grayscale for production, or pick Inferno or Turbo color maps to inspect depth at a glance. Style does not change the price.
Up to 10 minutes per clip
Process anything from a short social clip to a ten-minute sequence in one request, billed at 1 credit per second.
How it works
Upload your video
Add an MP4, WebM, or MOV up to 10 minutes. The duration is read locally to estimate credits before you start.
Pick a depth style
Choose Grayscale for generation and compositing, or Inferno / Turbo when you just want to preview depth visually.
Download the depth video
Processing runs in the cloud. When it finishes, preview the result, download it, or grab it later from your history.
Built for depth-driven workflows
Motion and camera replication
Extract depth from a reference clip and drive a depth-controlled video model to reproduce its camera path and movement with new content.
VFX and compositing
Use the depth pass for depth-of-field, fog, volumetric light, and depth-based color grading in post.
Relighting and 3D-aware edits
A stable depth pass makes relighting, parallax, and 2.5D effects predictable across the whole shot.
AR occlusion
Let virtual objects pass correctly behind real ones by feeding a consistent depth map into your AR pipeline.
FAQ
1 credit per second of input video. Duration is rounded up to the next second with a minimum charge of 4 credits, and a single clip can be up to 600 seconds (600 credits). The depth style does not affect the price.
Use Grayscale for any control or compositing workflow: white is closest to the camera and black is farthest. Inferno and Turbo are color visualizations meant for viewing only.
No. Depth is relative, normalized across the clip. It is ideal as a control signal or for compositing but is not a calibrated distance measurement.
Yes. The depth video keeps the input resolution and frame rate. Videos above 30 fps are returned at 30 fps with the same duration.