Image to Video Prompts: 40 Copy-Paste Examples for Any Photo

A good image to video prompt is one short line that names what moves in the photo and how: one subject motion, one gentle camera move, and the pace, with everything else left as it was photographed. "Slow wind through her hair, soft camera push-in" works. "Make it move" steers nothing, and "she stands up and walks to the window" asks the photo for a pose it does not have. Below are 40 copy-paste image to video prompts grouped by photo type, each with [BRACKETS] for the parts you fill in, plus spoken lines for talking photos, the prompts that fight the photo, and which clip length and aspect ratio to pick. They are written for MakeAIVideo's photo animation tool, but the structure works in any tool that animates a single still.

Written by Jamie Partridge, Founder of MakeAIVideo.

What makes a good image to video prompt?

An image to video prompt is not a scene description. The photo already is the scene, so the prompt only has to say what changes over the next 5 or 10 seconds. That is why the best image to video prompts are short, and why the motion box in the image to video tool is optional. Leave it blank and the motion is inferred from the photo, which suits most portraits and landscapes.

When you do want a say in it, build the line from five parts, in this order, and drop any you do not need:

1. Subject motion. What moves, and only one thing: a breath, a blink, hair in a breeze, steam, a tail wag.
2. Camera move. One gentle move or none: a slow push-in, a soft drift, a slow pan. Skip it when the subject motion is enough.
3. Atmosphere. The scene motion around the subject: clouds, water, rain, fog, light shifting across a wall.
4. Pace. "Slow", "subtle", "gentle". Small, slow motion keeps faces and products looking like themselves.
5. What to keep still. Name it when it matters: "the bottle stays still", "the face stays as it is".

Here is the anatomy in one line:

Subtle breath and a slow blink, light breeze in the hair, slow push-in, background stays still

That is real motion inside the frame, not a Ken Burns effect, which Wikipedia describes as "a type of panning and zooming effect used in film and video production from non-consecutive still images". A pan slides a camera over a still. An image to video prompt asks the photo itself to move.

Two more rules from the app itself. The motion note holds up to 500 characters, but one line almost always beats a paragraph, because several motions compete and blur into each other. And a line you want someone to say never goes in the motion prompt. It goes in its own box, "What should they say?", covered in the talking photo section below.

Copy the skeleton, not just the prompt. Every prompt below follows the same shape: one subject motion, one camera move at most, and a word for pace. When a result is not quite right, change one part of the line and generate again rather than rewriting all of it. Each new take from the same photo gives a fresh pattern of motion. Start 7 day free trial →

Image to video prompts for portraits and selfies

Portraits are the hardest case, because the face is the whole point and any drift in likeness shows. Keep the motion small. A breath, a blink and a little hair movement read as life; a big head turn reads as a different person. If the portrait is a character you plan to post every week rather than a one-off photo, build it once in the AI influencer generator so the same face comes back every time.

1. The living portrait (5 s, 9:16)

Subtle breath and a slow blink, gentle hair movement, camera holds still

2. Outdoor wind (10 s, 9:16)

Slow wind through [HER/HIS/THEIR] hair, soft camera push-in, [BACKGROUND] stays still

3. The half smile (5 s, 1:1)

A slow, small smile forms, eyes stay on the camera, no head movement

4. Golden hour (10 s, 9:16)

Warm light slowly shifts across the face, a soft blink, gentle camera drift to the [LEFT/RIGHT]

5. Profile picture loop (5 s, 1:1)

Subtle breath, slight sway, [CLOTHING] moves gently in a breeze, camera still

Image to video prompts for product photos

A single product on a clean background animates best. Busy flat-lays give the model too much to move at once, and rigid objects on a table mostly hold still, so the motion has to come from light, liquid, fabric or the camera. Keep the product itself still: a bottle that warps is worse than no motion at all. For a narrated product ad with several scenes, a single clip is one ingredient; build the full video with prompt to video.

6. Premium light sweep (5 s, 1:1)

Soft light shift across [PRODUCT], gentle camera drift, [PRODUCT] stays still

7. Slow orbit feel (5 s, 9:16)

Slow push-in on [PRODUCT], reflections move slowly across the [GLASS/METAL] surface

8. Fabric product (10 s, 9:16)

[GARMENT] moves gently in a light breeze, folds shift softly, camera holds still

9. Drink or skincare (5 s, 1:1)

Condensation slowly runs down [BOTTLE/GLASS], soft light flicker, product stays still

10. Listing hero (10 s, 16:9)

Slow pan across [PRODUCT] on [SURFACE], shadows shift slowly as if the sun is moving

Image to video prompts for landscapes, skies and water

Landscapes are among the most reliable photos to animate, because clouds and water move so cleanly. This is where 10 seconds earns its extra credits: the motion has room to land before the clip ends. The aurora example shows the type at its best, a single still of the northern lights over a mountain lake turned into a 10 second vertical clip where the aurora and sky move while the scene stays photoreal.

11. Lake and sky (10 s, 16:9)

Slow cloud drift, gentle ripple on the water, camera holds still

12. Coast (10 s, 9:16)

Waves crashing on [BEACH/ROCKS], slow push-in

13. Night sky (10 s, 9:16)

[AURORA/STARS] slowly move across the sky, water reflects the light, mountains stay still

14. Weather (5 s, 16:9)

Rain falling softly over [SCENE], light mist drifts low, slow camera drift

15. Forest or field (10 s, 16:9)

Light breeze moves through the [TREES/GRASS], sunlight flickers through the leaves, slow pan to the [LEFT/RIGHT]

Image to video prompts for pets

Pet photos work for tributes and social posts, and the motion should feel like the animal rather than a cartoon overlay. Keep the pet in the frame: a prompt that sends a dog running out of shot asks for movement the photo cannot supply. Ears, tails, eyes and fur are the parts that sell it. If you are making a series of pet clips for a feed, keep them all at the same ratio so they sit together in a grid or a Reel, and a Reels maker can wrap them into a narrated vertical video later.

16. The classic (5 s, 9:16)

Slow tail wag and an ear twitch, [PET] stays in frame

17. Sleepy cat (10 s, 1:1)

Slow breathing, a sleepy blink, whiskers twitch, camera holds still

18. Window light (10 s, 9:16)

[PET] slowly turns [ITS] head toward the window, fur moves softly in the light

19. Outdoor dog (5 s, 16:9)

Ears flap gently in the wind, tongue out, grass sways behind, slow push-in

Image to video prompts for old family photos

Old photos need respectful micro-motion: the scene breathes while faces, clothes and the room stay close to the original. Restore the print first, because a faded or scratched copy gives you a faded, scratched clip; the tool animates the photo you give it rather than repairing it. Colourise it too if you like. A colourised photo tends to read as a moment from life, a black-and-white one more as an effect. Use 10 seconds, and keep big head turns and broad smiles out of the prompt, because on an old portrait that is where it starts to look uncanny.

20. Grandparent portrait (10 s, 1:1)

A single breath, a slow blink, face and clothes stay as they are

21. Couple on the steps (10 s, 9:16)

Both breathe gently, a slight shared sway, the [HOUSE/STREET] behind stays still

22. Group photo (10 s, 16:9)

Everyone breathes softly, a light breeze moves [CLOTHES/HAIR], camera holds still

23. Memorial piece (10 s, 9:16)

Soft light slowly brightens on the face, a slow blink, very slow push-in

With several people in one photo, everyone is animated at once. Give one shared motion idea, like a gentle sway, rather than directions for each person. It works best when the faces are a similar size and similarly lit.

Image to video prompts for food and drink

Food is mostly about steam, pour, drip and light. These prompts suit menus, launch posts and short ads, and they work best on one dish shot fairly close with the plate filling the frame. A wide table spread has too many items competing for motion. For a narrated vertical video around it, a vertical video maker builds one from a single idea.

24. Hot dish (5 s, 1:1)

Steam rises slowly from [DISH], soft light flicker, plate stays still

25. Coffee (5 s, 9:16)

Steam curls up from the cup, a slow swirl in the crema, slow push-in

26. Cold drink (5 s, 9:16)

Bubbles rise in [DRINK], ice shifts slightly, condensation slides down the glass

27. Sauce or syrup (10 s, 1:1)

[SAUCE] slowly drips over the edge of [FOOD], camera holds still

Image to video prompts for architecture and interiors

Buildings and rooms are rigid, so the motion comes from the camera, the light and anything soft in the frame: curtains, plants, water, people far in the background. This is the type for property listings and venue posts. Agents can see where a short clip fits alongside other property marketing video options.

28. Living room (10 s, 16:9)

Slow push-in through the room, curtains move gently, sunlight shifts on the floor

29. Exterior at dusk (10 s, 16:9)

Clouds drift slowly behind [BUILDING], windows glow a little brighter, camera holds still

30. Skyline (10 s, 16:9)

Slow pan across the skyline, lights flicker on in a few windows

31. Pool or garden (5 s, 9:16)

Gentle ripple on the pool, palm leaves sway, slow camera drift

Image to video prompts for art, illustrations and album covers

Paintings, digital art and album covers animate well, and at 5 seconds the style has less time to drift. Ask for motion that already belongs to the image: if the cover has neon, let it flicker; if the painting has a sea, let it move. Prompts that ask the art to change style pull it away from what you made.

32. Neon cover (5 s, 1:1)

Neon lights flickering, slow push-in

33. Painted landscape (10 s, 16:9)

The painted clouds drift slowly, brushstroke texture stays as it is

34. Character illustration (5 s, 9:16)

[CHARACTER] blinks slowly, hair and cape move gently in a breeze, background stays still

35. Abstract art (5 s, 1:1)

The shapes flow slowly in one direction, colours stay as they are, camera still

One photo, one motion, one ratio. Every render makes one clip at the aspect ratio you picked, so choose 9:16, 16:9 or 1:1 before you generate. The motion is framed around that ratio, and re-cropping afterwards can cut out the part that moved. Start 7 day free trial →

Image to video prompts for talking photos: spoken lines that work

A photo can also speak. Type a line under "What should they say?" and pick a voice. If the image shows a face, the person says it with an AI voice and lip sync. If it does not, as with a product or a landscape, the line plays as a voiceover. Keep the motion prompt small when you add a line, because the mouth and face are already doing the work. For a presenter reading a full script rather than one line, use a talking avatar instead.

Length matters more than wording. As a guide, about 150 words makes a 60 second talking head, so a 10 second clip holds roughly 25 words and a 5 second clip about half that. Paste a line into a speaking-time checker to see if it fits. Front-load the line, because the first seconds decide whether anyone keeps watching; a hook writer gives you openers to pick from.

Each prompt below is two parts: the motion note, then the line in brackets for the second box.

36. Product intro

Motion: subtle breath, slow push-in
Say: "Hi, I'm [NAME]. This is [PRODUCT], and it does one thing really well: [BENEFIT]."

37. Launch teaser

Motion: a slow blink, slight smile
Say: "Something new is coming on [DATE]. Follow so you don't miss it."

38. Event invite

Motion: gentle hair movement, camera holds still
Say: "Join us at [EVENT] on [DATE]. Link in the bio to save your spot."

39. Voiceover on a landscape (no face)

Motion: slow cloud drift, gentle ripple on the water
Say: "This is [PLACE] at [TIME OF DAY]. Here's why it's worth the trip."

40. Character post

Motion: subtle breath, slow blink
Say: "Quick tip for [AUDIENCE]: [ONE-SENTENCE TIP]. Save this for later."

Making a real person appear to say something is the line to be careful with. YouTube's rules on altered or synthetic content require creators to disclose AI content that "makes a real person appear to say or do something they didn't do", and TikTok's AI-generated content label exists because its policy "requires people to label AI-generated content that contains realistic images, audio or video". Meta set out the same direction when it said it would require a disclosure and label on photorealistic video or realistic-sounding audio that was digitally created or altered. Get permission for any likeness that is not yours, and say the clip was made with AI.

Image to video prompts to avoid: what fights the photo

Most bad results come from prompts that ask the photo for something it does not contain. The tool animates the image you uploaded rather than generating a new scene, so a prompt that pulls against the photo loses.

• Big pose changes. "She stands up and walks to the window" will not happen. The subject keeps the pose from the photo and moves within it. If you want someone mid-stride, upload a photo where they are already walking. To put your character into real moving footage instead, character swap is built for that.
• New objects or people. "A dog runs into the frame" or "add a birthday cake" asks for things that are not in the photo. Start from a photo that already has them, or build a new scene with a text prompt instead.
• Text. Text in the image tends to warp once it moves, so do not prompt for moving titles or signs, and avoid photos with dense text overlays. Add captions in the edit, not in the image.
• Several motions at once. "Wind in the trees on a stormy day with rain and lightning" is four ideas. They compete and blur together. Pick the one that matters.
• Nothing at all. "Make it move" or "animate this" names no motion, so it steers nothing. Either leave the box blank and let the motion be inferred, or name one motion.
• Big expressions on old portraits. Broad smiles and head turns on an old photo are where it tips into uncanny. Stay with a breath and a blink.

When something goes wrong, the fix is often in the source photo. Faces that smear point to a small, compressed or filtered photo, so upload a sharper copy. A background that swims means a busy scene with no clear subject, so crop tighter. The minimum is 300 by 300 pixels, but a sharp photo gives the model far more to work with.

Image to video prompt settings: 5 vs 10 seconds and aspect ratio

The prompt is half the result; the length and ratio are the other half. One photo becomes a 5 or 10 second 1080p clip in 9:16, 16:9 or 1:1, with one ratio per render. In the app, 5 seconds is labelled Fastest and 10 seconds Smoother motion.

The app shows the exact cost before you generate. Plans start at $9 a month, and every plan starts with a 7-day free trial; the full breakdown is on the plans page.

For ratio, match the place the clip will live:

• 9:16 vertical for TikTok, Instagram Reels and YouTube Shorts. Sprout Social's video specs guide recommends exporting short-form vertical placements "like Reels, TikTok, Shorts and Stories" at 9:16, and lists YouTube Shorts as 9:16 or 1:1.
• 16:9 landscape for YouTube, websites and listing pages. The same guide puts landscape YouTube video at 16:9.
• 1:1 square for the Instagram feed, Facebook and product grids.

If you are not sure, start vertical. HubSpot's video marketing statistics report that short-form videos like TikTok, Instagram Reels and YouTube Shorts are the most commonly used video formats among marketers. Need the same photo in two shapes? Render it again at the other ratio; each crop is generated separately so the motion fits the frame.

Test at 5 seconds, ship at 10. A 5 second take is the cheaper way to check that a prompt moves the right thing. Once the motion is right, render the 10 second version for anything slow, like a landscape or a memorial portrait. Start 7 day free trial →

From image to video prompts to a longer video

A single clip is the right size for a feed post, a listing or a profile loop, and the wrong size for a 30 second narrated video. A narrated multi-scene video can be built from one written idea, and if you are scripting it yourself, our guide on how to write an AI video script covers the structure.

If your photos are of one recurring character rather than real people, the prompt library approach scales further. Our 50 scene prompts for AI influencers use the same fill-in-the-brackets format for a character who posts every week, and the character generator keeps the face consistent from post to post.

Image to video prompts FAQ

What is an image to video prompt?

An image to video prompt is a short line telling the tool what should move in your photo and how. The photo already sets the scene, so the prompt only describes the motion: one subject motion, a gentle camera move, and the pace, like "slow wind through her hair, soft camera push-in".

Do I need a prompt to turn an image into a video?

No, the motion prompt is optional. Leave it blank and the motion is inferred from the photo, which suits most portraits and landscapes. Write a prompt when you want to steer it, for example to keep a product still while the light moves.

How long should an image to video prompt be?

One line is usually enough. The motion note holds up to 500 characters, but one subject motion and one camera move beat a paragraph, because several motions compete and blur together. If a result misses, change one part of the line and generate again.

Can the person in my photo walk, dance or change pose?

Not reliably. The subject keeps the pose from your photo and moves within it: a breath, a blink, hair in a breeze, a change of expression. If you want someone mid-stride or mid-dance, upload a photo where they are already in that pose.

How do I make a photo talk?

Type the line under "What should they say?" and pick a voice. If the photo shows a face, the person says it with lip sync; on a photo with no face, it plays as a voiceover. A 10 second clip holds roughly 25 words, and a 10 second clip with lip sync is about 216 credits.

Should I pick 5 or 10 seconds?

Pick 5 seconds for product loops, food and feed posts, and 10 seconds when the motion needs room to land, such as a landscape, an interior or a memorial portrait. A silent 5 second clip is about 92 credits and a 10 second clip about 163.

Which aspect ratio should I use for an image to video clip?

Use 9:16 for TikTok, Reels and Shorts, 16:9 for YouTube and websites, and 1:1 for the Instagram feed and Facebook. Choose it before you generate, because the motion is framed around the ratio and re-cropping later can cut out the part that moved.