Add your characters, objects, and scenes as references, then describe how they come together.
Upload the people, products, or other materials you want to use, then describe the scene. Reference to video AI uses @ mentions to tell which reference belongs to each character or object. Depending on the model, you can also add video or audio to guide movement and sound.
Show this AI video maker what you want to use instead of describing every detail from scratch. Add your references, assign their roles, and explain what happens.
Upload character photos and describe a scene between them. Decide who speaks, who reacts, and what happens next. Use the results to develop short dramas or test a scene before filming.
Use a supported video reference to show the movement you have in mind. Combined with character images, it works like an artificial intelligence animation generator for dance clips, reactions, and other short-form content.
Upload a product photo and a model reference, then describe how they appear together. Show someone holding a lipstick, presenting a skincare bottle, or using a product in a daily routine.
Reference to video comes down to three steps: upload, assign, and generate.
Add the images, video clips, or audio supported by your selected model. Choose clear materials that show the people, products, or movements you want to use.
Use @ mentions to connect each reference to your instructions. Say who each person is, which object they use, and what they should do.
Choose your video settings and generate. Check the characters, products, and movement, then download the result or revise your prompt.
AI picture to video starts from one image. With references, you can put several characters, a product, and a movement into the same scene.
Upload character photos and describe a meeting, an argument, or a surprise reveal. Add a key object as another reference to give the characters something to interact with.
Combine a model reference with a lipstick, skincare, or perfume photo. Describe how the model holds or presents the product to create a short beauty ad.
Pair character images with a supported motion reference to guide a dance, martial arts sequence, or action scene. Clearly assign each movement to the right person.
Use a singer’s photo to create a stage or studio performance. Describe gestures, lighting, and camera movement, and add audio where the model supports it.
See how creators use AI reference to video to show the characters, products, and movements they want.
It feels like a videographer AI that already knows my cast. I upload the character photos and spend the prompt on what they do, not on how everyone looks.
As an AI video creator for our shop, it lets us put our product image with a model reference and try a scene before arranging a shoot.
Video reference to video AI made sense once I tried it: a short clip shows the movement I mean much better than a long written description.
I can put the characters and a key object into one scene. It makes the story easier to discuss with the team.
Plain pic to video AI only animates one photo. Here the references keep a particular person and product in the scene, which is what I needed.
I had tried an AI video generator free plan elsewhere, but it could not keep my singer consistent. With a singer reference I can try stage lighting and camera moves first.
Find answers about supported references, @ mentions, character consistency, and how closely the AI can follow your source materials.
It is a way to generate videos using uploaded references and a text prompt. References can guide character appearance, products, movement, or sound, depending on the model. Some people find it by searching “Vidu AI reference to video”; the workflow here is the same idea.
Image to video AI generally uses an image as the starting frame. Reference to Video uses uploaded materials to guide a new scene, which may include several characters, objects, or reference types.
If you are wondering how to reference a video, upload the clip as a reference, select it with @, and say what it should guide, for example: “@Woman copies the dance from @Video.” The same works for images: “@Woman picks up @Product and shows it to @Man.”
You can upload references and draft the prompt for free. Generating uses credits, and reference to video does not use the free attempts some image and text models offer. If you searched “reference to video AI free”, credits from the free credits page can be spent here.
Yes, where the model supports multiple references. Give each person a clear role and say who performs each action. This helps reduce confusion between characters.
A supported model can use the video to guide motion, but poses, timing, camera movement, and the background may change. It does not guarantee a frame-by-frame copy.