13 Credits
NSFW, violent, and misleading content is prohibited. Content Policy
Examples
Same Character, New Scene
The character from the reference images walks through a sunlit forest path, looking up at the trees, gentle breeze moving her hair, slow tracking shot, warm afternoon light
Product in Context
The sneaker from the reference image on a wet city street at night, a runner steps into frame and the camera pushes in on the shoe, neon reflections, cinematic lighting
Mascot Story Beat
The mascot from the reference images waves at the camera from a festival stage, confetti falling, crowd cheering in the background, bright daylight, handheld shot
Features
- Upload one to three reference images (JPG, PNG, WEBP up to 10MB)
- Keeps a character's, product's or mascot's appearance consistent
- Describe any new scene, action and camera move in text
- Choose Veo 3.1 (flat 13 credits) or Seedance 2.0 (per second, native audio)
- Landscape and vertical outputs
Powered by Leading AI Models
Veo 3.1
Google
Seedance 2.0
ByteDance
100K+
Videos Generated
15K+
Happy Users
4.7 ★
Average Rating
Frequently Asked Questions
What is reference-to-video generation?▾
Reference-to-video generates a new video from one to three reference images plus a text prompt. The images show who or what should appear (a person, a product, a mascot), and the prompt describes the scene, action and camera. The model keeps the subject's appearance consistent in the new clip.
Which AI models support reference-to-video?▾
On Kensa, reference-to-video runs on Veo 3.1 by Google DeepMind and Seedance 2.0 by ByteDance. Veo 3.1 is a flat 13 credits per clip. Seedance 2.0 is billed per second, supports clips up to 15 seconds and includes native audio at no extra cost.
What can I upload as a reference?▾
Images only: JPG, PNG or WEBP, up to 10MB each, one to three per generation. Video and audio files are not accepted as references. Clear, well-lit images with the subject fully visible give the most consistent results.
How is this different from image-to-video?▾
Image-to-video animates your image as the first frame, so the clip starts from that exact picture. Reference-to-video uses your images as a guide for the subject's appearance and generates a new scene around it, so the first frame does not have to match any of your images.
How many credits does reference-to-video cost?▾
With Veo 3.1, each clip costs 13 credits. With Seedance 2.0, the cost depends on length and resolution, starting at about 7.7 credits per second at 480p. The exact cost is shown before you generate.