Describe your scene, pick the cinematography, and get a prompt tuned for your model — Seedance 2.5, Veo 3.1, Sora 2, Kling 3.0 or Runway Gen-4.5. Free, instant, runs in your browser.
Your prompt will appear here…
How to write better AI video prompts
Every strong video prompt answers five questions: who/what is in the frame, where they are, what happens during the clip, how the camera behaves, and what it looks like (style, lighting, mood). This generator assembles those five layers in the order each model parses best.
Model-specific quirks
Seedance 2.5 — responds well to structured prompts: subject → action → camera → style. Set duration and resolution in the UI, not in the prompt. Supports up to 50 reference images for character consistency.
Veo 3.1 — prefers flowing, natural cinematic sentences, like describing a film shot to a director. Audio cues ("waves crashing softly") work well.
Sora 2 — handles long, detailed paragraphs and complex physics. Describe cause and effect explicitly.
Kling 3.0 — strong at human motion; be explicit about body movement and facial expression.
Runway Gen-4.5 — keep it concise; short, punchy visual descriptions outperform long paragraphs.
FAQ
Is this prompt generator free?
Yes — completely free, no signup, no watermark. It runs entirely in your browser; nothing you type is sent to any server.
Should I mention resolution or duration in the prompt?
For most modern models (Seedance 2.5, Veo 3.1, Sora 2), no — set resolution, duration and aspect ratio in the generation interface or API parameters. Keep the prompt about the content of the shot.
Why do the generated prompts differ between models?
Each model was trained on differently formatted caption data. Seedance parses structured comma-separated blocks well, while Veo and Sora respond better to natural film-language sentences. The generator reorders and rephrases the same creative choices per model.