seedance-2-0

Generate cinematic video clips with synchronized audio using ByteDance Seedance 2.0 APIs.

Updated Sep 1, 2026
One-click install
npx skills add https://github.com/zamansepeti43/c-rak-agent --skill seedance-2-0-zamansepeti43
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: seedance-2-0
Source: https://github.com/zamansepeti43/c-rak-agent/tree/main/video-engine/.agents/skills/seedance-2-0
Command: npx skills add https://github.com/zamansepeti43/c-rak-agent --skill seedance-2-0-zamansepeti43

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Producing cinematic trailers, teasers, and multi-shot video clips with synchronized audio requires coordinating multiple AI video gateways, prompt structures, and provider-specific parameters, which is error-prone and inconsistent without a unified workflow. ## Core Features & Use Cases - Multi-gateway video generation: Route Seedance 2.0 requests through fal.ai, Replicate, Runway, or Higgsfield via the OpenMontage video_selector with automatic provider scoring and fallback. - Structured prompt methodology: Apply the Higgsfield prompt format with shot-structure openers, temporal markers, identity-lock language, and camera negations for multi-shot and lip-synced generations. - Reference-conditioned generation: Use up to 9 images, 3 video clips, and 3 audio clips to maintain character identity across shots with anti-drift fallback strategies. - Use Case: Generate a 10-second 21:9 cinematic trailer clip with multi-shot cuts, quoted dialogue lip-sync, and native sound effects by calling video_selector with preferred_provider set to seedance. ## Quick Start Ask the agent to generate a 10-second cinematic trailer clip with Seedance using a multi-shot prompt and save it to the project assets folder.

Frequently Asked Questions about seedance-2-0

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a video with Seedance 2.0 in OpenMontage?

Call the video_selector tool with preferred_provider set to seedance, plus prompt, operation, aspect_ratio, duration, and output_path. The scoring engine picks whichever configured gateway (fal.ai, Replicate, Runway, Higgsfield) is available.

What prompt structure works best for Seedance 2.0 multi-shot videos?

Open with a shot-structure declaration such as the montage or single-POV opener templates, then add environment detail, character blocks, beat-by-beat choreography with temporal markers, and explicit camera negations like no cuts or no zoom.

Which API keys does Seedance 2.0 support across providers?

Seedance 2.0 works with FAL_KEY for fal.ai, REPLICATE_API_TOKEN for Replicate, RUNWAY_API_KEY for Runway Enterprise non-US plans, and HIGGSFIELD_API_KEY with secret for Higgsfield. Any one configured gateway is sufficient.

How do I keep character identity consistent across Seedance shots?

Use the reference-to-video endpoint with a clear front-facing portrait, reuse the same reference across all shots, and stack identity-lock phrases like the same character, no drift, and do not alter clothing category or primary color.

When should I use the fast variant instead of standard Seedance 2.0?

Use the fast variant only for b-roll, previews, or latency-constrained drafts at lower cost. Avoid it for slow-motion, multi-shot, or complex camera moves, since it routinely misses those on the first attempt.

What are the limitations of Seedance 2.0 video generation?

Clips cap at 15 seconds and 1080p on some endpoints, readable text rendering is unreliable, dialogue over six words risks lip-sync drift, and complex music scoring should be handled by dedicated audio tools instead.