Direct with references, not words alone.
Upload images, clips, and audio to show the model which people, products, movement, style, and sound the new video should follow.
Image, video, and audio guidance · Built for visual consistency
Image, video, and audio
Input types
1 studio
Full workflow
4K
Max output
Add your references
Turn your references into a brand-new video
Add images, videos, or audio, then explain what each reference should control in the generated clip.
Click to upload images, videos, or audio
jpeg, png, webp, mp4, mov, mp3, wav
Generate multiple videos in one batch with the same settings, up to 10 videos.
Guide with uploaded media
Control identity, motion, style, and sound with references
Use images, videos, and audio to guide identity, movement, visual style, and sound more clearly than text alone.
01 / Identity guidance
Keep characters and products recognizable
Use image references to ground appearance, wardrobe, shapes, materials, and key details across a newly generated performance or product moment.
- Reference people, characters, products, or props
- Call out specific media directly in the prompt
- Preserve the details most important to recognition
02 / Motion reference
Borrow motion without copying the whole shot
Use a video reference to communicate timing, camera energy, body movement, or performance nuance that would be difficult to explain precisely with words alone.
- Guide rhythm, gesture, and camera behavior
- Combine movement from one source with identity from another
- Translate complex performance notes into visual context
03 / Multiple reference types
Guide image, motion, and sound together
Combine visual references with audio cues, then use the prompt to explain how each uploaded file should affect the new video.
- Reference images, videos, and audio in one workflow
- Create more specific scenes from less ambiguous input
- Keep every instruction and reference in one generation
04 / Combined references
Use several references in one new shot
Assign a clear role to every uploaded asset, then let the model synthesize identity, movement, style, and sound into a single new result.
- Give each reference one explicit creative purpose
- Resolve complex direction with visible examples
- Create a new composition rather than a simple copy
How it works
Create a video from your references
Upload the media you want the model to follow and explain the role of each file.
Add your references
Upload images for identity and style, clips for motion, and audio for sound or performance context.
Mention what each reference controls
Use reference chips in the prompt to connect each file to a person, product, action, camera move, or cue.
Generate and evaluate
Choose the model and output settings, create the shot, and refine any instruction that needs more emphasis.
When details must stay consistent
Consistency across people, products, and motion
Use references when text alone cannot clearly describe a person, product, movement, style, or sound.
Character continuity
Keep a recognizable character present across new scenes, actions, and environments.
Product storytelling
Ground shape, material, and brand details while generating a new product performance.
Performance transfer
Use an example clip to guide movement, timing, gesture, or camera energy.
Creative adaptation
Combine an approved visual world with new action, pacing, and audio direction.
Family story films
Use familiar faces and treasured photos to shape a short film around a shared memory.
Celebration tributes
Combine photos, video, and audio cues into personal birthday, wedding, or anniversary messages.
Travel diaries
Guide a new sequence with real trip photos while keeping places and people recognizable.
Hobby demonstrations
Use example movement and reference images to visualize crafts, recipes, and personal projects.
Creator stories
How creators get more consistent results
“We can point to the exact product, the exact movement, and the exact visual world. The prompt no longer has to carry everything.”
“Reference clips make choreography notes much more useful. We show the rhythm we mean and build a completely new shot around it.”
“Character consistency is what unlocks narrative work for us. The image references give every new scene a recognizable anchor.”
FAQ
Reference to video, explained
How images, video clips, and audio can guide a new generated video.
Create with references
Ready to Experience Multi-Modal AI Video Creation?
Upload the people, products, movement, style, or audio you want to keep consistent, then generate a new clip.
Create with references