AI Video From a Photo: How It Works, Templates, Prices

What "AI video from a photo" means

An AI video from a photo is a short clip that an image-to-video model generates from a still picture. You give it a photo of a person, a pet or a couple, and the model invents motion, camera movement, a scene and sound around them.

Zebra (the AI photo editor at zebra.tg, not Zebra Technologies) wraps this into templates. Instead of writing a prompt and hoping for the best, you pick a scene that is already tuned, like "Ride your giant pet" or "100 years of fashion", and only add your photos. The website uses the same templates as the Zebra iOS app.

Here is a real result from one photo of a person and one photo of a cat:

Key facts at a glance

Templates you can use

There were 21 templates in September 2026. The catalog is updated on the server, so new ones appear without an app update. Price in coins is listed for each; it matches the video length in seconds.

Pets

• Ride your giant pet (5 coins): you ride your cat or dog, grown to giant size, through a city you choose. This is the "giant cat in New York" trend.
• A giant hug from your pet (5 coins): your pet, now huge, wraps you in a hug. This is the "AI hug video" trend, with your own pet.
• Giant pet on the sofa (7 coins): your giant pet lies down on you like the warmest blanket ever.
• Become your pet (5 coins): one jump, and the owner turns into their pet.
• My cat got a job (5 coins): your cat at work, in a job you type in.
• I'm already asleep (5 coins): your pet caught scrolling a phone past bedtime.
• Your pet's daily commute (5 coins): your pet riding the metro.
• A giant in the city (8 coins): a giant version of your pet walks through a city.
• City guardian (8 coins): your pet protects the city from a monster.

Solo

• Wall Street moment (5 coins): your moment of triumph, Wall Street style.
• Meet your other selves (7 coins): an elevator full of different versions of you.
• 100 years of fashion (10 coins): five eras of style in one clip.
• Five sides of you (10 coins): five different looks, still recognizably you.
• Balcony vacation (5 coins): a pool on your balcony.
• A year outside your window (7 coins): four seasons through one window.
• Beyond the clouds (6 coins): you fly above the clouds over your city.

Couples

• Our journey to success (8 coins): from walking together to driving a sports car.
• Hollywood sunset (7 coins): a villa, a sports car and a sunset.
• Us in the 90s (6 coins): the two of you in a 90s-style video.
• Together in every lifetime (8 coins): you find each other in different eras.
• A moonwalk for two (6 coins): a walk on the Moon together.

How to make an AI video from a photo in 4 steps

1. Pick a template. Open /ai-video and tap a template to watch its example video.
2. Add your photos. Each template asks for one or two photo groups (for example "Person" and "Your pet"), and each group takes 1-3 photos of the same person or pet. Some templates also ask for a city or a job.
3. Choose a format and start. Pick 9:16, 4:5, 1:1 or 16:9 and confirm the price in coins. You can close the tab while the video renders.
4. Download the video. It appears on the page and in your profile, usually in 2-5 minutes. Download it within 3 days.

You can run up to 4 videos at once and up to 30 videos in 24 hours.

Tips for photos that work

• Use a clear, well-lit photo with the face visible. Sunglasses, heavy shadows and blurry shots give the model less to work with.
• Add 2-3 photos of the same person or pet when you can. More angles help the model keep the likeness.
• Keep one subject per group. A group photo in the "Person" slot confuses the model about who should be in the video.
• Pets: show the whole animal. A photo where the fur color and shape are easy to see works better than a close-up of the nose.
• Avoid tiny images. Photos that are too small are rejected, so use the original file, not a thumbnail.
• Pick the format for where you will post. 9:16 for Stories, Reels and TikTok, 4:5 for an Instagram post, 16:9 for YouTube.

Honest limits

• AI results vary. The same photos can give a great clip one time and a strange hand or face the next. The model invents the motion, so small details are not always accurate.
• Templates only, no free prompts. You get tuned scenes, not a blank text box. That is faster, but less flexible than a general video model.
• 720p, 5-10 seconds. Enough for Stories and Reels, not for a big screen.
• 3-day storage. Finished videos are deleted from the server 3 days after completion. Download the ones you want to keep.
• Your photos go to an AI provider. Generation runs on an external AI video model, so the photos you add are sent to it. Some photos or texts can be rejected by the provider; if generation fails, the coins are returned.
• Sign-in required. Videos and coins are tied to your account so you can find them in your profile.

How it compares to general AI video tools

If you want a still image instead of a video, the AI Art Generator restyles one photo for 0.5 coin; see how to turn a photo into AI art.

FAQ

Ready to try? Pick a template at /ai-video.