Add reference images, video, or audio, then guide the motion with your prompt.
Upload a few images, then type @ in the prompt to pick them: @Image1 for the person, @Image2 for the outfit or product, @Image3 for the location. Reference to video AI reads each tag and keeps every element in the same clip, so the face, the clothes and the scene stay the way your pictures show them.
Three references, one prompt. The model, the black silk dress and the balcony each come from their own @ tag, and the clip keeps all three intact.
Prompt: @Image1 wears @Image2 on the balcony in @Image3, turns slowly, a strap slips off her shoulder, ends on a sultry smile.
A pet photo, a skateboard shot and a skatepark. Tag them, and the dog rides that exact board into that exact bowl.
Prompt: @Image1 rides @Image2 into the bowl at @Image3, wobbles out at the top, hops off and grins at the camera.
An artificial intelligence animation generator that starts from your own pictures, not random stock looks.
Tag a model photo and a swimsuit or dress, and get a flirty lookbook clip on a yacht, a beach or a rooftop. The face and the outfit stay true to your images.
Plain image to video AI starts from one frame. Here you hand the model several references and tell it, with @, what each one is for.
Tag a face, a dress and a room, and the shot is cast, dressed and set the way you planned. Think of it as a videographer AI that never swaps your look for its own.
Generate Video from Images
Use the same @Image1 across many prompts and your mascot or model stays the same in every clip. Ideal for series, ads and storytelling.
Direct the Motion
Reference images, video clips and audio together. Tag a sneaker, a dance move and a beat, and get one clip that uses all three, where the model supports it.
Explore the Controls
I sell swimwear. With plain image to video the bikini kept changing color between clips. Now I tag the model and the product separately, and the swimsuit looks exactly like my catalog photo.
Our brand fox used to look different in every AI clip. I keep one @Image1 for him and only change the scene. The whole series finally feels like one character.
I tagged my model, a black silk dress and a Paris balcony photo. The clip looked like we flew to Paris for the shoot. My clients could not tell.
Type @, pick the file, keep writing. No guessing which picture the model will use. It is the clearest way I have found to do reference to video.
I tagged my cat and a tiny chef hat for fun. The video went around my whole family chat. Free to try, and it took minutes.
Wondering how to reference a video idea from your own files? Three steps.
Add the images, and if the model allows, video or audio clips.
Type @ in the prompt and pick a file. Say what each one does: who, what, where.
The AI video maker renders the clip in minutes. Preview it and save.
Quick answers about references, @ tags and pricing.
It is a way to make a video from several reference files at once. You upload images, tag them with @ in the prompt, and the AI uses each one for the part you assign: a person, a product, a location or a style.
Upload your files, then type @ in the prompt box and pick one from the menu. It appears as a tag like @Image1. Write what it should do, for example "@Image1 wears @Image2 in @Image3".
Yes, on models that support it. Video reference to video AI lets you tag a clip for motion, camera move or style next to your images. The upload panel shows which file types the selected model accepts.
Image to video animates one picture from its first frame. AI reference to video combines several references into a new scene, so you can put your character in a new outfit and a new place in a single clip.
You can start free and try reference to video without paying. Longer or higher resolution clips use credits, and the exact cost is shown before you generate.
It depends on the model. Most accept several images, and some also take video and audio. The upload panel shows the live limit for the model you picked.
That is the point of reference to video. Use a clear, well-lit photo for @Image1, and the AI video creator keeps the same face across shots and across new clips.
Only with their permission. Do not make intimate or misleading clips of a real person without consent. We remove that content and may close the account.
