Images to video AI: several pictures in one clip, each in the shot you choose - RefVid
Turn your own photos into videos in minutes · Your first video is free, no watermark
First video free, no cardGenerating confirms everyone pictured agreed. Terms
Yes, RefVid turns several images into one AI video. Add up to 7 pictures of a place, a look or a style; each is named @Image1, @Image2 and so on in the order you add it, and typing @ in a shot puts one there.
The pictures go to the model in that order, on Kling 4, Kling 3.0 Omni and the other models. A saved character or product can join them as its own @tag, or the clip is made from the pictures and your words alone.
2 pictures, no saved character
How images to video AI works on RefVid
- You add pictures to the studio's composer, and each one becomes a tag in the order you add it: @Image1, @Image2, up to @Image7. Writing a tag in a shot tells the model which picture that shot is about. When you press Generate, every picture goes to the model as a reference image, in that order, with the shots that name them.
- A reference picture is not a first frame. The model reads it as the place, the look or the style and draws a new scene from it, with its own camera move. In the clips above, shot 1 shows the courtyard from @Image1 redrawn as a moving shot, not the photo itself made to wobble. When you do want a clip to open or end on an exact picture, the picture's menu offers that as well (below).
Multiple images to video AI: pictures, elements and @ tags
- A clip can mix two kinds of reference. A picture is one image for this clip: a room, a street, a product shot, a frame from your storyboard, a colour mood. An element is a character or product you saved once from 1 to 5 photos, written as its name (@Mia) and reused in every clip. One clip takes up to 7 pictures and up to 3 elements.
- Each picture's menu has four choices: Use in this shot, which writes its tag where the cursor is; Start the video on this picture; End the video on this picture; and Remove image. Kling 4 and Kling 3.0 Omni take both a first and a last frame; the other models take a first frame as a reference the clip opens on, and no last frame. On Kling 3.0 Omni pictures and elements share one limit, so every element you add leaves room for one picture fewer. When a set doesn't fit the model you picked, the studio says so before any credits are held.
Step by step: several pictures, several shots
- Open the studio, then click + Image, or drop or paste pictures anywhere on the composer. Each picture takes its tile and its number at once and uploads in the background. Add + Shot for each beat of the clip, up to 3 shots; a new shot starts at 3 seconds and you set each shot's length.
- In each shot, type @ to open the list of your pictures and elements, pick one with the arrow keys and Enter or Tab, and write what happens. The head of each shot shows what it uses, with an × to take a reference out of that shot and a + to add a new picture straight into it; dropping a picture on a shot does the same. Then pick the model, the shape and sound, and press Generate.
- Stuck on the words? Improve reads the shot's pictures and turns a short idea into a full shot, keeping every tag you wrote. With nothing written, it writes the shot from the pictures alone. It describes what moves and how the camera goes, not what the picture already shows.
The exact prompts behind the two clips
- The three-shot clip on Kling 3.0 Omni, 3 seconds per shot. Shot 1: "Low camera glides through the sunlit courtyard garden in @Image1 , warm haze drifting." Shot 2: "Slow push-in on the playful scene from @Image2". Shot 3: "The camera pulls back as golden light fades over the courtyard." Shot 3 names no picture, and the model kept the courtyard from shot 1.
- The 5-second clip on Kling 4, one prompt: "The camera glides from the courtyard garden in @Image1 toward the playful scene from @Image2 , soft afternoon light." Both prompts are short on purpose. The pictures carry the place, the characters and the drawing style, so the words only say where the camera goes and how the light feels.
Kling multi image to video and the other models
- Every model in the studio takes several pictures: Kling 4, Kling 3.0 Omni and Seedance 2.5, Wan 3.0, Vidu Q4, Gemini Omni Flash, PixVerse C1, and Happy Horse 1.1. You always write @Image1, @Image2; RefVid renames the tags into each model's own way of naming its images before the request goes out, and sends the pictures in the same order, so the second picture is still the one you called @Image2.
- The models differ in how many images they take in total, counting the photos of your elements, and in what else they accept: lengths run up to 30 seconds on Kling 4 and up to 15 on Kling 3.0 Omni. If you need more pictures than a model allows with your elements, remove a picture or pick another model; nothing is charged for a request the studio refuses.
What it costs, and checks before you post
- Pictures add nothing to the price. A clip costs by model, length, resolution and sound, the same with one picture, seven or none, and a new account's first 5-second video on Kling 3.0 Omni is free. RefVid adds no watermark, and a clip that fails gives its credits back. Current prices live on the pricing page.
- A picture is redrawn, not pasted: a person or product that only appears in a picture can change from shot to shot. When a face or a label has to stay the same across clips, save it as an element instead. Upload only pictures you have the right to use, watch the clip at full size before posting, and label realistic AI video as AI-generated where the platform asks for it.
Start from a template
Each one opens the studio with the shots, shape and length already set. Add your own product and generate.
More how-to guides
Frequently asked questions
Can I make an AI video from multiple images without a character?
Yes. Add the pictures, write a shot that names them with @Image1, @Image2 and so on, and press Generate; no saved character or product is needed. Both clips on this page were made that way, from two pictures and a few words.
How many images can I use in one AI video?
Up to 7 pictures per clip, plus up to 3 saved elements. Some models count pictures and element photos against one limit (Kling 3.0 Omni among them), and the studio tells you before generating when a set doesn't fit.
What does @Image1 mean in a prompt?
It is the first picture you added to the clip; @Image2 is the second, and so on. Writing the tag in a shot tells the model that this shot uses that picture. Type @ in a shot to pick one from a list instead of typing it.
Can each shot use a different picture?
Yes. Each shot names its own tags, and each shot's head shows what it uses. In the three-shot clip above, shot 1 uses @Image1 and shot 2 uses @Image2. A clip has up to 3 shots.
Can a picture be the first or last frame of the video?
Yes. A picture's menu offers Start the video on this picture and End the video on this picture. Kling 4 and Kling 3.0 Omni take both; the other models take the first one as a reference the clip opens on.
Does adding more pictures cost more credits?
No. The price depends on the model, the length, the resolution and sound, not on how many pictures the clip uses.
By Louis Chi