Animate a photo with AI
Attach a picture and describe the motion — AI turns a landscape, painting or pet photo into a video clip up to 12 seconds long, with sound.
See what AI creates
Made by our users in minutes — no designers, stock photos or long retouching.
How to animate a photo in 4 steps
A landscape, painting, pet, product or an old photo of a place. The sharper the shot and the clearer the main subject, the cleaner the motion.
What comes alive and what the camera does: “clouds drift, ripples on the water, slow push-in”. Or pick a direction and the prompt fills in.
Landscape 16:9 or vertical 9:16, 2 to 12 seconds, quality and sound. Nothing is generated until you press Send in the Studio.
The clip lands in the Studio chat with no visible watermark — ready for social media, websites and messengers.
Part of the Zerocoder platform
One tool of a platform where AI generates websites, images, video, text and documents — under one subscription.
Animating a photo with AI turns a single still into a short video: water and clouds start moving, the camera glides in and ambient sound is added. It runs in the Zerocoder Studio on the Seedance model, so you can turn a photo into video and animate a picture online without editing software. Common jobs have their own pages: logo animation and product video. For a clip from a text description alone, with no photo, use the AI video generator.
Photo to video AI: which pictures come alive best
Shots where the scene itself suggests motion work best: water, sky, smoke, foliage, fire, fabric in the wind. A lake landscape becomes a clip with ripples and drifting clouds; a still life gets a flickering candle and steam over a cup. Paintings and illustrations work well too — the model tries to keep the brushwork while adding motion inside the scene. The camera orbits a product as highlights slide across it; a pet blinks and turns its head. No portrait of your pet yet? Create one with the AI pet portrait tool first, then bring it to life here.
How to write a prompt that animates a photo
The working formula: what moves + what stays still + what the camera does + what you hear. For example: “clouds drift slowly, the water ripples, the mountains and the house stay still, static camera, wind ambience, no speech”. Open with “keep the composition and details as in the photo” — the model uses your picture as a reference and may otherwise invent things. One or two motions per clip look more natural than five at once. If the first take misses, change one phrase and compare two versions.
Bring an old photo to life: what works
“Bring an old photo to life” often means restoration. A faded scan can go through image enhancement first — it normalizes levels and adds sharpness, though it won't remove scratches — then animate it here. Old photos of places without people work best: a hometown street, a country cottage, a railway station, a courtyard. If there are passers-by in the frame, the model may reject the picture. In the clip, clouds drift and trees sway, and the film grain and black-and-white tones usually stay. Family portraits won't work: the model rejects pictures of real people, as the FAQ below explains.
What to do with the finished clip
The clip downloads as an MP4. Turn it into a GIF for a chat, cut extra seconds with video trimming and shrink it for messengers with video compression. Longer stories are assembled from several clips in any video editor: animate a few photos with similar prompts and stitch them together. Add captions and titles in the editor — the model may distort letters inside a generated video. Transparent backgrounds aren't available.
Reviews
No reviews yet — be the first.
More in Free tools
View category →
From the blog
All posts →
FAQ
Which AI animates the photo?
Seedance, a video model that turns a still into a clip with camera and scene motion. Open the Studio from this page and it's already selected. The Studio accepts up to 4 reference photos per clip, but one is enough to animate a specific picture.
Can I animate a photo for free?
Sign-up is free and new users get free starter credits. Video uses far more credits than images, though: starter credits cover at most a short draft, and a full clip needs a one-day trial or a Studio plan. The Studio shows the exact cost before you send.
Can I animate a photo of a person?
As a rule, no: Seedance rejects photos of real people — that's the model provider's requirement — and your credits come back. The tool doesn't make talking portraits, lip sync or face swaps. Please don't upload photos of children.
What kind of photo works best?
A sharp shot with a clear main subject and something that can move on its own: water, sky, leaves, smoke, fabric. JPG and PNG work. On this page you can attach a photo up to 3.5 MB; attach larger files in the Studio — your prompt carries over.
Will the photo stay exactly the same?
The model uses your picture as a reference: composition, colors and key details usually hold, but small elements can shift and text may warp. Add “keep the composition and details as in the photo” to the prompt and try a couple of takes.
How long is the video?
2 to 12 seconds per generation, 5 by default. Longer clips and higher quality cost more credits. A longer video is assembled from several clips in any video editor — animate a few photos and stitch them together.
Can I make a vertical video?
Yes: the Studio defaults to 16:9, and vertical 9:16 for Reels, TikTok and Shorts is one click away. Pick the format closest to your photo's proportions so the composition changes less.
Is there sound?
Yes, by default the model adds ambient sound and effects: wind, water, city hum. Describe in the prompt what should be heard, or switch sound off in the settings. Speech is the model's weak spot, so don't count on a voiceover.
What if the clip fails?
If a generation doesn't start, breaks during rendering or hangs, the credits are returned automatically. Credits aren't refunded for a result you simply didn't like — refine the prompt: name one main motion and ask to keep the photo's details.