Models / Seedance / Reference to video

Seedance reference to video prompting guide

Last checked 2026-10-11Seedance 2.x (ByteDance) · Seedance 2.0 / 2.0 Fast / 2.5

How to prompt Seedance in reference-to-video mode. Every point links its source; community tips are marked. Your agent gets the same guide through the Atlas MCP (get_guide).

Text to videoImage to videoStart & end frameReference to videoVideo to videoLip sync

Prompt structure

asset bindings→subject with ref→action→scene→lighting color→camera→style→constraints

What the text should describe

What not to describe

Camera vocabulary

Responds well to

Duration and limits

Multi-shot

Avoid and failure modes

Negative prompt

Not supported; describe what you want instead. source ↗

Real prompts, credited

One reference per job (body vs face) · @ZentrixHQ · original ↗
AI video hack nobody's talking about: give each reference image a job.

Seedance 2.5 on @makeugc 

Don't upload one picture and hope the model keeps the character. A single image has to carry the face, the outfit and the props at once, and something starts to drift by the third shot.

Upload two instead, and assign them in the very first line of the prompt:

"Use image 1 for her full body, dress and broom. Use image 2 for her face, eyes and hair."

One image for the body. One image for the face. One sentence that says which is which.

Then describe the outfit with the exact same words every time it matters, and end the prompt with "Preserve her face, hair, outfit and broom throughout."

The result: a 25-second chase where the witch who hits the road sign is still clearly the same witch who falls into the corn.

🔖I’ve pinned the prompt in the comments - save it so you don’t lose it.
Two characters from @[char refs] · @aimikoda · original ↗
Seedance 2.5 Prompt:
Animate @[char1 ref]  as Neri, 24, and @[char2 ref]  as her father Daro, 58. Preserve each reference's face, hair, outfit and proportions.

Visual style: animated painted forms with crisp contours and faceted light. Keep the references' long silhouettes, angular faces and broad clothing shapes, with restrained brush texture and expressive anime-influenced acting. Extend this treatment to the folded dragon skin and monumental stone tiers. Amber light cuts through teal dusk. The girl keeps youthful proportions.

30 seconds in a sixty-meter circular open-air arena with steep stone tiers and packed, continuous rows of spectators. A girl in a moss-green cap stands between seated adults in the east front row, immediately behind the arena's continuous guardrail. Her left hand grips the rail; her right holds one ivory paper flower with folded petals and a rolled stem. The tw
@Image 1 defines four kittens and the street · @aivoxyy · original ↗
SEEDANCE 2.5 PROMPT — 15s, 16:9
@Image 1 defines the four kittens, their accessories, their order, and the street. Preserve exactly: grey kitten with pink bow and katana in front, orange kitten with brown bob wig and toy submachine gun second, beige kitten with gold tiara third, dark grey kitten with green helmet and rocket launcher last. The wall, palm, sidewalk and black sedan stay fixed in place.

[GENERATION GOAL]
One continuous 15-second photorealistic comedy shot of four real kittens walking upright like an elite movie squad heading into a mission, then panicking at a cucumber. Real live-action look, real fur, soft overcast daylight. No cuts, no music, no subtitles, no slow motion.

[CAMERA]
Smooth sideways tracking shot at kitten height, moving RIGHT at their walking speed, keeping the whole squad centred. At 9 seconds the camera stops moving and holds.

[0-3s] Empty sidewalk besi

Next

All Seedance modesSeedance model, prices & shotsFrom $0.16/s · pricesKling reference to videoVeo reference to videoHailuo (MiniMax) reference to videoWan reference to video

Sources

Step 1 of 4

Which tool do you use?