Direct Seedance 3D Animation
SkillAI & modelsPlan and prompt a stylized 3D animated short for Seedance 2.0 with stable character identities, environment continuity, causal story beats, shot-specific performance, camera, sound, and clip handoffs. Use when a story needs one 4–15 second generation or a longer multi-clip plan; not for photoreal live action, a single still, or generic prompt polishing.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Direct Seedance 3D Animation skill
What this skill tells your AI
The instructions your AI receives, as published by t8mars/minimax-h3-prompt-skill-t8 in skills/direct-seedance-3d-animation/SKILL.md and read by ahel’s review.
Turn a story idea into Seedance-native clip prompts. This is a T8-authored companion to the MiniMax repository-bundled 3d-animation-short-generator; use the upstream Skill for its H3 workflow.
Workflow
- Lock aspect ratio, total duration, dialogue language, and whether the deliverable is one clip or a multi-clip short.
- Write a one-sentence premise and a causal beat spine. The protagonist must cause or respond to each important change.
- Define each recurring character with a stable label, two or three identity anchors, wardrobe, silhouette, and emotional baseline.
- Define environment anchors and screen geography before camera moves.
- Split a long film into clips of at most 15 seconds. Give every clip a visible opening state, change, payoff, and handoff frame/action.
- Choose generation assets deliberately: character images for identity, scene images for environment, and videos only for a named action/camera/style role.
- Compile each clip with
references/template.md, then audit continuity across clips.
Seedance rules
- Use one compact paragraph for one continuous event; use consecutive
镜头Nblocks for multi-event clips. - Never put exact per-shot seconds inside a Seedance prompt. Keep assembly timing in the external shot plan.
- Bind every ambiguous subject to
@图片Nor a declared video role; never leave an unused reference. - Specify performance through gaze, posture, breath, face tension, anticipation, follow-through, and recovery.
- Use one primary camera behavior per shot and preserve direction across handoffs.
- Put dialogue in
{...}, sound in<...>, and music in(...)only when required. - Do not emit H3 fields, H3 labels,
[Shot N], or timestamp alignment.
Boundaries
Do not call a paid model or upload assets without a separate request. Do not import this official-companion entry into ComfyUI; the user's node already includes the upstream presets.
Deliver
Return the locked story/identity/environment sheet, clip list, one Seedance prompt per clip, cross-clip handoff notes, and the final continuity audit.
Signals
- GitHub stars
- 217
- Forks
- 21
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
direct-seedance-3d-animation- Source
- github.com/t8mars/minimax-h3-prompt-skill-t8