Duration = source video length (max ~15s) Β· Token cost scales with video duration
Upload File
Paste URL
Choose video
Uploadingβ¦
β Video ready
π¬ Multi-shotπ€ Character Consistencyπ΅ Audio Referenceπ₯ Video Referenceβ¨ High Realism
Train a character once from a portrait photo. After training completes, copy the
@omni-character:<id> token and paste it into any Omni prompt for consistent identity.
At least 1 reference image is required. Use @image1β¦@image9, @video1β¦@video3, @audio1β¦@audio3 in your prompt to reference the uploaded media. Trained characters: @omni-character:<id>.
Est. Costβ500500 tokens Β· VIP Standard
π Frame-to-Frame MotionπΌ 1 or 2 Imagesβ‘ Fast & VIP Options
Upload 1 image (first frame only) or 2 images (first + last frame). The model interpolates motion between them guided by your prompt.
VIP Fast auto-infers aspect ratio from your reference images.
Est. Costβ625625 tokens Β· Global Standard
π Merge Clipsπ¬ Preserve Orderπ Custom Aspectβ± 5β60s per clip
Upload your clips in order. Each clip must be 5β60 seconds. Minimum 2 clips required. Use Auto aspect ratio to preserve the first clip's dimensions.
Combiner is a low-cost utility β 2 tokens per combine.
Estimated Costβ250250 tokens Β· Global Standard
β οΈ This generation uses β make sure your prompt is ready before submitting.
Cinematic Prompt Agent
powered by KAMOD AI
Describe any video idea β get a production-ready Seedance 2.0 prompt back, structured like a film director wrote it.
Reference images (optional Β· up to 3 Β· character, location, style)
οΌ
οΌ
οΌ
Gemini will analyse character appearance, clothing, and location from these images.