Reference to Video Generator
Create Reference to Video clips with AI by uploading reference images or reference videos, writing a prompt, and choosing models built for consistent subject, style, motion, and scene guidance.
Reference to Video AI · Reference to video · Reference image to video · Reference video input · Subject consistency
Sample Video
Gallery
Your Reference to Video history
Review Reference to Video tasks created from this browser or account. Preview finished videos, reopen prompts, share links, download files, or delete items you no longer need.
Why it works
Reference to Video AI for consistent creative control
Use Reference to Video when text alone is not enough. Upload reference images or reference videos, then guide the result with prompt, model, format, audio, and advanced settings.
Reference-driven generation
Use reference images or reference videos to guide subject identity, scene style, composition, and motion.
Reference-capable models only
The model picker on this page only shows AI video models that support Reference to Video workflows.
Format and audio controls
Tune aspect ratio, duration, resolution, generated audio, uploaded audio, or source audio when the selected model supports it.
Built for reusable visual direction
Create campaign clips, product videos, character shots, style tests, and branded scenes from reusable reference media.
Workflow
How to create Reference to Video AI clips
Upload reference media
Add reference images, reference videos, or style images supported by the selected Reference to Video model.
Write a Reference to Video prompt
Describe how the AI should use the references, including subject consistency, camera movement, scene mood, and output format.
Generate and reuse
Submit the Reference to Video task, preview the result, then reuse the prompt or references for the next iteration.
Use cases
Use Reference to Video for consistent AI video assets
Resources
Reference to Video guides and related tools
Explore related AI video workflows for turning reference media, images, and prompts into finished short videos.
Image to Video AI
Use image-to-video when you have one source image and need a direct animated clip.
Open guideText to Video AI
Use text-to-video when you want to create a video from prompt alone without reference media.
Open guideAI Video Gallery
Browse examples and prompts before creating a new Reference to Video task.
Open guideImage Generator
Create reference images first, then use them in the Reference to Video generator.
Open guideUser stories
Reference to Video for creators and teams
“Reference to Video helps me keep the same character and visual style across campaign clips.”
“I can upload product references and get video ideas that still feel connected to the original brand assets.”
“The filtered model picker makes Reference to Video much easier because I only see models that support references.”
FAQ
Reference to Video questions
What is Reference to Video?
Reference to Video uses reference images or reference videos to guide an AI video result, helping preserve subject identity, visual style, motion, or scene direction.
Which models appear on this Reference to Video page?
The model picker is filtered to models that support the reference+text mode, such as reference image or reference video workflows.
Can I upload both reference images and reference videos?
Some Reference to Video models support reference images, some support reference videos, and some support both. The upload controls change based on the selected model.
Does Reference to Video support audio?
Audio options depend on the selected model. Some Reference to Video models support generated audio, uploaded audio, or keeping source audio.
When should I use Reference to Video instead of Text to Video?
Use Reference to Video when you need stronger control over subject identity, product appearance, brand style, source motion, or reusable visual direction.
Can I use Reference to Video for commercial projects?
You can use suitable outputs commercially when you have rights to the reference media and review the generated result before publishing.