Choosing a video model for prompting (FAQ)
Use the in-app motion prompt coach or prompt upsample for model-specific rules. This page is only a high-level chooser — not the full prompting rulebook.
Video model prompting — quick comparison
Use the in-app motion prompt coach or prompt upsample for model-specific rules. This page is only a high-level chooser — not the full prompting rulebook.
Seedance 2.0
Best when you have image + video + audio references and want reference-first control (motion, camera, pacing from refs). In Ciaro, cite refs with {{ref:…}}, {{vref:…}}, {{aref:…}} chips in the prompt.
Kling V3 Omni
Best for character locking, multi-shot sequences, Elements, motion-reference video, and native audio/voice references in production workflows.
Runway Gen-4.5
Best for production shots with strong image references and visual continuity — prompt describes action and camera; references define look.
Runway Aleph 2.0
Best for editing existing footage (replace, relight, restyle) — prompt states what changes and what must stay.
HappyHorse 1.1
Best for director-style shot prompts: clear subject, visible action, camera terms, and audio described in text. Supports image references only (up to 9) — not video/audio reference slots. Start frame or reference images, not both.
Ray / Veo / Sora
Ray: camera-first motion. Veo: physical cause-and-effect. Sora: narrative emotional beats. Each has its own coach rules when selected in the timeline.
Not sure?
Open the motion prompt field with your model selected and ask the prompt assistant to rewrite your idea — it loads the full guide for that model automatically.