Result
Your result will show up here.
Depth Video Generator
Create a depth video from a clip for free. Use its motion and scene structure as a reference alongside an image and prompt in compatible AI video workflows.
How to Generate a Depth Video
A browser tool for short clips
Upload a Short Video
Upload a video up to 30 seconds and 200 MB. On first use, the browser downloads the depth model.
Choose the Look
Choose a frame rate, grayscale or Inferno, and an available export format.
Download the Depth Video
Preview and download the depth video, or use the Depth Video node in Canvas to connect it to a compatible workflow.
Turn a video into a depth map
Each frame estimates relative distance and turns the scene into a depth map. A moving subject or camera produces a sequence of maps that can help a compatible model follow the source clip’s spatial layout and motion. Depth is estimated, not a precise 3D reconstruction.
Guide a new take on the original scene
For example, turn a walking clip into a grayscale depth video, then pair it with a new character image and a prompt describing a different setting. A compatible model can use the sequence as visual guidance while creating a new interpretation; the result will not reproduce every movement exactly.
Create in your browser or on Canvas
Use the free browser form for a downloadable depth video, or connect the Depth Video node to a compatible video input in Canvas. Need the sound from the same clip? Extract its audio separately and use it where reference audio is supported.
Frequently Asked Questions
What is a depth video?
A depth video shows estimated relative distance in each frame rather than the original colors. It can provide spatial and motion guidance for compatible video generation workflows.
How do I use a depth video for AI video generation?
Create a depth video from a source clip, then supply it to a model that supports depth or video conditioning, together with your reference image and prompt. Support and results vary by model.
Which frame rate should I choose?
Start with the default 12 fps for a balance of detail and processing time. Choose 8 fps for a quicker draft, or try 15 or 24 fps for smoother motion. Higher frame rates generally take longer; also check your destination model’s requirements.
Should I choose grayscale or Inferno?
Choose grayscale when a model expects a conventional depth map, and check its input requirements. Inferno uses a color ramp that can make depth easier to inspect visually; it may not be accepted as a depth input.
Should I export WebM or MP4?
WebM is the default. Choose MP4 when your browser supports H.264 encoding and the receiving tool needs MP4. Check the destination model’s accepted formats.
What are the video limits?
The standalone tool accepts clips up to 30 seconds and 200 MB. Output contains no audio. Videos larger than about 512 pixels on their longest edge are scaled down; smaller videos are not enlarged.
Will a depth video recreate the original video exactly?
No. The depth map is an estimate of relative scene depth, and the final result also depends on the model, prompt and reference image. Use a clear source clip with visible subjects and camera movement, and expect to experiment with the prompt and model settings.
Is it free and does it upload my video?
Depth generation is free and uses no credits. On this standalone page, processing runs in your browser; Canvas may upload assets and outputs for connected nodes. Other AI generation services may charge separately.
Why does the first run take longer?
The browser downloads the depth model on first use and may reuse it later. Processing speed depends on your device and browser. Use a recent desktop Chrome or Edge with WebGPU and WebCodecs.
Generate a depth video from a short clip
Free tool. Use the form or connect the node in Pictell Canvas.