Depth Map (Video)

Turns a video into a temporally stable depth-map video (Video Depth Anything). Feed the grayscale result to Seedance 2.5 as a reference video to transfer only the motion, camera and timing, without the face, clothes or room. Priced per second of video.

Try Depth Map (Video) in the playground

Last updated View as MarkdownAgent setup

Depth Map (Video)

Node type: video:depth_anything_video
Category: video

Description

Depth Map (Video) turns a video into a temporally stable depth-map video (Video Depth Anything) for motion-only transfer: wire your source clip into Video and a grayscale depth video comes out of Depth Video. The shapes, body motion and camera move survive; the face, clothes, colors and room do not. Wire that output into Seedance 2.5's Reference Video port next to your character images, and say in the prompt that the clip is a depth map to take only the motion, timing and camera from. Keep Colormap on grayscale when the result feeds another model. Seedance takes its length from the reference clip, so trim the source first; one run takes up to 2400 frames (80 seconds at 30 fps). It costs about 5 credits per second of the source video.

Pricing

  • Cost: ~5 credits per second
  • Per-second rate: The figure is the rate for a 5-second clip. Shorter clips can cost a bit more per second. You're always charged for the real length.
  • Billing: An estimate for a default run. You are charged for what the run actually produces, settled when it finishes.

Canvas ports

These appear as port handles on the left side of the node.

ID Label Details
video_url Video VIDEO (required). Source clip. Up to 2400 frames (80s at 30fps). Trim it first if it feeds Seedance 2.5, which takes its length from the reference.

These render as form fields in the right-side config panel when the node is selected.

ID Label Details
model Model TEXT · options: VDA-Small, VDA-Base, VDA-Large · default: "VDA-Large"
colormap Colormap TEXT · options: grayscale, turbo, inferno, magma, viridis · default: "grayscale". Keep grayscale when feeding Seedance or another video model; color maps leak color into the take.
resolution Resolution TEXT · options: auto, 360p, 480p, 720p, 1080p · default: "auto". Auto keeps the input resolution (max 1080p).
side_by_side Side by Side BOOLEAN · default: false. Output the original and the depth map side by side in one frame. For showing the effect, not for feeding a model.

Outputs

ID Label Type
video Depth Video VIDEO

What it is for

Motion-only transfer. Wire any clip in and you get a grayscale depth video back: the shapes, body motion and camera move survive, the face, clothes, colors and room do not. Wire that output into Seedance 2.5's Reference Video port next to your character images, and the model has nothing left to copy but the motion. Tell it so in the prompt ("@Video1 is a depth map, take only the motion, timing and camera from it").

Keep Colormap on grayscale when the result feeds another model; the color maps are for viewing. Side by Side puts the original and the depth next to each other, which is handy for showing the effect but not for feeding a model. Seedance takes its length from the reference clip, so trim the source first.

What it cannot do

The depth video holds only the source clip's shapes and motion, so it gives Seedance no motion for anything your stills add. In one run, a crowd added in the stills stayed frozen while the person moved; that run's prompt also said "no invented action", so drop that line when your stills add people. When the new scene needs its own motion, start from a source clip filmed somewhere like it.

Cost

Priced per second of the source video. Up to 2400 frames per run (80 seconds at 30 fps).


Auto-generated from the Wireflow node registry.

For AI agents: the full documentation index is at /llms.txt, and most docs pages are available as Markdown by adding .md to the URL.

© 2026 Wireflow. All rights reserved.