Seedance

Seedance

by ByteDance

AI video generation model producing multi-shot cinematic videos from text prompts and mixed media inputs

See https://seedance.ai

Features

• Multi-shot cinematic output — generates story-driven videos with consistent characters across shots
• Synchronized audio — lip sync support in multiple languages; audio track generation
• Multimodal inputs — accepts up to 9 images, 3 video clips, and 3 audio files as reference materials
• Reference-based control — use @image1 syntax to specify main character; @video2 for camera movement; @audio3 for tempo
• Camera replication — copy camera moves or choreography from reference clips
• 2K resolution output — up to 2K, multiple aspect ratios (16:9, 9:16, 4:3, 21:9, 1:1)
• 4–15 second clips — configurable duration per segment
• HeyGen integration — powers HeyGen’s Avatar V and digital twin cinematic workflow
• n8n workflow compatible — automate generation pipelines via n8n nodes

Superpowers

Seedance 2.0 addresses the biggest limitation of AI video: consistency across shots. By letting you specify character references, camera references, and audio references separately with @tag syntax, you maintain visual continuity when combining multiple generated segments into a longer piece. The HeyGen integration adds avatar/digital twin support on top, turning a single prompt into a full production pipeline: script → avatar → b-roll → scene editing. For content creators producing social video at scale, this reduces production overhead dramatically. Competes directly with Kling, Wan 2.5, and Veo 3.

Pricing

• Access via Seedance platform (credit-based, check current tiers)
• Available through HeyGen for avatar-integrated workflows