Animate a photo
Subtle-motion prompts: breathing portraits, wind in the hair, slow push-ins, living landscapes.
AI Video Pro · Prompt Library
AI Video Prompts That Work: A Library From Real Generations
Not invented examples. These are the prompt patterns that produce finished, downloaded clips in AI Video Pro most often, cleaned up and grouped by what you are trying to make.
A good image-to-video prompt is short and specific. It names what moves, what stays still, how the camera behaves and what the light is doing. Almost every failed generation is missing one of those four, and almost every reused prompt has all four.
This page is the index. The twelve prompts below are the general-purpose patterns; each use-case library linked underneath goes deeper into one kind of clip.
How it works
- 1
Start with the subject and the frame
"In a 9:16 vertical frame, the person in the photo…" tells the engine what it is animating and for which format.
- 2
Add motion, camera, light
One movement for the subject, one for the camera, one lighting note. That is a complete prompt.
- 3
Lock what must not change
"Keep the same face, hairstyle, outfit and background as the photo." Say it every time; it is a constraint, not decoration.
What people make with it
Cinematic camera moves
Dolly-ins, orbits, Time Freeze, slow motion, golden hour and anamorphic night looks.
Hug videos
Two adults from one photo, one warm embrace, consent-first.
Walking videos
Walk toward the camera, down a street, along a beach, across a runway.
Extend a video
Continuation prompts that keep the same subject, light and motion across the join.
Prompt library
12 prompts distilled from the generations that get finished and downloaded. Copy one, or open it in the generator with the photo you want to animate.
The complete template
[Frame: 9:16 vertical.] [Subject: the person in the photo] [Motion: slowly turns their head toward the camera and smiles.] [Camera: still.] [Light: soft window light.] Keep the same face, hairstyle, outfit and background as the photo.
Fill in the brackets; delete the labels.
Identity lock
Keep the same face, hairstyle, outfit, body proportions and background as the photo. No new people, no new objects.
Append to any prompt.
Minimal motion
Subtle movements, slight shifts, slow pace, stationary camera, smooth motion, consistent identity, stable background.
Camera-only motion
The subject holds still while the camera slowly pushes in over the full clip. Natural, believable motion, cinematic lighting, no cuts.
One continuous take
One continuous shot, no cuts, no transitions, no camera-angle changes. The action plays out in real time from start to finish.
Weather as motion
A gentle breeze moves the hair and the fabric; leaves drift across the frame. The person and the camera stay still.
Product turn
The product rotates slowly a quarter turn and settles centered. Soft studio lighting, consistent reflections, sharp label, black background.
Two people, one scene
The person on the left steps toward the person on the right and they share a warm hug, holding for a few seconds. Faces stay exactly as in the photo. Camera still.
Walk to camera
The person walks toward the camera with slow, confident steps and stops close, looking into the lens. Same face and outfit as the photo. Camera still.
Timelapse
Timelapse: clouds race, shadows sweep across the ground, the buildings stay perfectly still. Wide static shot, high contrast.
Continuation
Continue the exact same scene and motion. The subject keeps doing what they were doing at the same pace. Same lighting, same camera, no cuts.
Spoken line
The person looks directly at the camera and says, in a warm natural voice: "[your line here]". Lips match the words, small hand gesture, camera still, soft light.
Use with an AI voice from the voice library; adults only.
How to structure an image-to-video prompt
Subject, motion, camera, light, lock. Five parts, one sentence each at most. The order matters less than the presence: a prompt with all five and no adjectives beats a paragraph of mood words with none.
Write in the present tense and describe the clip as if it already exists ("the camera pushes in", not "please make the camera push in"). Name durations when they matter ("hold for two seconds"). Put the frame format first if you are making vertical video.
What to leave out
Style tags ("4K, masterpiece, trending") add nothing to a video engine. Brand names, real people and anything that changes who the person is do not belong in a prompt and are checked automatically. GoCrazyAI is for adults; the people in your photos must be adults who have agreed to appear.
Frequently asked questions
- Where do these prompts come from?
- From the prompt patterns that most often lead to finished, downloaded clips in AI Video Pro, rewritten for clarity. No individual user prompt is published.
- Do the same prompts work for text-to-video?
- Mostly. Drop the "as in the photo" lock and describe the subject in one sentence instead. The camera, motion and light language is identical.
- How long should a prompt be?
- Two to four sentences. Longer prompts usually add contradictions, and the engine resolves contradictions by ignoring one of them.
- Who can appear in the videos?
- GoCrazyAI is for adults (18+). Use your own photos, or photos of adults who have agreed to appear. Uploads and prompts are checked automatically before generation.
- Are the results marked as AI-generated?
- Yes. Every clip is AI-generated and carries provenance metadata.
- What does it cost?
- Generation is paid with credits or a subscription. The exact credit cost is shown on the Generate button before you confirm.
Ready to try it?
Open AI Video Pro, upload a photo and paste a prompt from this page.
Last updated: