Seedance 2.5 3D blockout camera control: clay renders as a reference
Seed describes clay-render referencing, a textureless 3D model that blocks camera and pose. Sume has no clay mode; you send the render as a video reference.

For 3D blockout camera control with Seedance 2.5, render a plain clay-style pass of your scene from the camera path you want and send it as a video reference. Sume has no separate clay-render mode; the docs list video references for the Seedance 2.x models, and you say in the prompt that the clip is a camera and pose guide.
ByteDance's launch post describes clay-render referencing, a textureless 3D model used to block camera and pose. What follows separates that vendor description from what Sume's API takes.
What is clay-render referencing?
Per the launch post, it lets a creator block camera and pose with a textureless 3D model rather than describe them in words. The post does not give Sume-specific settings, and this page does not claim any result quality; it explains how to send such a clip.
How do I send a blockout render on Sume?
Host the render at a public HTTPS URL and add it to input_references as a video_url entry on seedance-2.5. Per the Video generation docs, audio and video references are honored by the Seedance 2.x models. Your prompt then carries everything the clay clip cannot: the materials, the light, the subject and the mood.
What limits apply to the request?
Images, videos and audio in input_references count together, and the API refuses a request over 12 files. The model itself takes 4 to 30 seconds per request.
| Limit | Value |
|---|---|
| Length of the generated clip | 4 to 30 seconds |
Files in input_references | At most 12 in total |
| Reference type for the clay clip | video_url |
Will the result follow the blockout exactly?
Do not assume it will. Sume's docs describe references as guidance for a new generation, not as exact frames, and the launch post gives no accuracy figure. Run a short 480p pass first, compare the camera path and poses with your render, and adjust the prompt before you pay for a longer or higher-resolution take.
A useful check is to compare the framing at the first and last second with the blockout before you look at anything else.
How should I prepare the blockout clip?
Keep the render simple: one neutral material, the camera move you want, and figures posed the way the shot needs them. Detail in a clay pass competes with the prompt, so leave texture and lighting decisions to the text. Match the aspect ratio of the render to the aspect ratio you request, and keep the clip near the length of the shot you want.
The camera-move reference guide covers how a reference clip changes the estimate, and reference limits lists the vendor's own caps.
Sources
Related posts
More in Models
- Seedance 2.5 aspect ratios: 9:16 vertical, 21:9 and the rest
Seedance 2.5 on Sume accepts 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. Set aspect_ratio explicitly; 3:2 is rejected. Ratios, errors and the price effect.
- Seedance 2.5 audio reference: how to send an audio clip
Seedance 2.5 accepts an audio clip as a reference, even without an image or video. The audio_url request on Sume, the vendor's file limits and what to expect.
- Seedance 2.5 image to video: first frame and last frame
Send a first_frame, or a first_frame and last_frame, with seedance-2.5 on Sume. Rules for frame_images, how it differs from input_references and what to set.
- Seedance 2.5 omni reference: how @image, @video and @audio tags work
Seedance 2.5 prompts point at references with tags such as @Image 1. Sume rewrites lowercase tags like @image1 to the casing the model expects. Examples.
Written by Sume