Seedance 2.0 Fast. 4 to 15 seconds. This is the default render.
Models
Models behind the lock.
Auto renders on Seedance 2.0 Fast. Named Seedance models are there when the plan includes them. A script model writes the line. The reference set, not the model name, is what keeps the character.
Seedance 2.5
ByteDance via Runway
Named cinematic model with native voice, up to 30 seconds. Auto does not use this engine. Auto is Seedance 2.0 Fast.
Text to videoNative audioLip sync9:16 vertical4 to 15 s
ScriptsDeepSeek V4 Pro
DeepSeek via OpenRouter
Writes the hook, the spoken line, the scene, and the caption for the length you set.
Chat completionsJSON modeShort-form scriptsCaptions
CharactersClickPop Bible 1
ClickPop
Turns your description into a character bible. The reference photos are the lock that keeps the face and outfit.
Character bibleVoice specPersonaHandle and name
How a render is chosen
Auto first. Lock always.
Front, three-quarter, profile, and close-ups, created from the character you describe.
480p, 720p, and 1080p are frame sizes. Seedance routes in the studio currently deliver 720p.