How to animate a still photo with AI: 5 steps and prompts
Quick answer
To animate a still photo, upload it to an image-to-video tool, describe one clear movement, generate a short clip, and inspect the entire result before downloading. Start with subtle motion such as a slow camera push, blinking, or wind moving hair. If the person must speak recorded audio, use a talking-photo workflow instead.
Animate one still image
Upload a clear source image, describe one subtle movement, and inspect the full clip for changed faces, objects, text, and background details.
Try Image-to-VideoChoose the right photo-animation workflow
Image-to-video: add camera or subject motion to a still image.
Talking photo: animate a portrait to follow uploaded speech, song, or laughter.
Photo restoration and colorization: repair an old still before adding generated motion.
Video-to-video: restyle footage that already moves; it does not start from a still image.
These workflows can be combined, but evaluate one transformation at a time. Keep the original image and compare each intermediate output so you can identify where a face, product, logo, or background changed.
How to animate a still photo in five steps
Choose a clear source image. Use the highest-quality original available, with the main subject large enough to inspect.
Open Magic Hour Image-to-Video and upload the image. The current guest tool works in a browser without signup.
Describe one movement. Specify either subject motion, camera motion, or a simple environmental change before combining them.
Generate a short clip. Available duration, resolution, aspect ratio, audio, and model choices vary; check the selected settings before generating.
Review the full result. Inspect identity, hands, products, text, logos, background geometry, and the transition from the first frame to the last.
Open Magic Hour Image-to-Video. The current public page lists three guest generations per day, three seconds each, with 480p watermarked exports. Paid models and output options differ, so rely on the settings shown in the live tool rather than assuming every model has the same limits.
Three motion prompts to copy
The subject blinks once and gives a slight natural smile. Subtle breathing. Locked camera. Preserve facial identity, hair, clothing, and background.
Workflow recipe
Live-action cinematic scene. A realistic human character standing on a film set during a short break between takes. The character initially looks slightly sideways (diagonal gaze, not directly at the camera), then naturally turns their head a bit while speaking 1–2 short sentences in a relaxed, conversational tone. After speaking, the character gives a gentle, warm smile. Natural idle movements: casually adjusting hair, straightening clothing, lightly extending one hand forward as if explaining or reacting. Body language feels unposed and spontaneous, like an actor resting between shots. Background: a busy film crew behind the character — camera operators adjusting equipment, crew members walking back and forth, assistants carrying gear, continuous subtle motion. Camera remains mostly stable with a slight handheld feel, shallow depth of field, cinematic lighting, realistic skin texture, natural facial expressions. No exaggerated gestures, no stiff acting, smooth logical motion throughout the scene.
- Model
- veo3.1
- Resolution
- 720p
- Duration
- 6 seconds
- Frame rate
- 24 fps
Slow camera push toward the product. Keep the package, label, shape, colors, and surface unchanged. Soft background light moves slightly. No new text.
Slow camera pan to the right. Gentle wind moves leaves and grass. Preserve the buildings, horizon, lighting direction, and composition.
A prompt states the goal; it does not guarantee preservation. If a required detail changes, shorten the clip or simplify the movement before adding more instructions.
When to use Talking Photo instead
Use Magic Hour Talking Photo when a portrait must speak or sing along with uploaded audio. It starts from a photo and an audio file, then generates mouth and facial movement. Review lip timing, eyes, head motion, teeth, and facial identity across the complete clip.
Talking Photo and image-to-video answer different briefs. A general motion prompt is appropriate for atmosphere or camera movement; uploaded audio is the correct control when the speech itself must drive the animation.
How to animate an old family photo
Work from a copy of the best scan. If the image has scratches or fading, use Old Photo Restoration first. If the source is black and white, use Photo Colorizer only when predicted color is appropriate; generated color is not historical evidence. Compare faces, jewelry, uniforms, lettering, and background details after each step.
For a respectful result, start with minimal movement. Do not present generated speech or motion as an authentic historical recording, and get permission before animating a private person’s image.
A practical quality checklist
The face and body remain recognizable from start to finish.
Hands, teeth, eyes, clothing, and hair do not deform or flicker.
Products, packaging, logos, and written text remain accurate where required.
The camera movement has one clear direction and does not create impossible geometry.
The background does not add, remove, or merge important objects.
The aspect ratio fits the final placement and keeps the subject clear on a phone.
You have rights to the source image and the intended use of the output.
Frequently asked questions
Magic Hour’s current Image-to-Video page lists three guest generations per day without signup. Guest outputs are short, 480p, and watermarked. Confirm the live allowance before starting because product limits can change.
Start with one subtle action and a locked camera: “The subject blinks once and gives a slight natural smile. Subtle breathing. Preserve facial identity, hair, clothing, and background.” Add more motion only after the first result is stable.
Some models may generate facial motion or audio, but use a talking-photo tool when uploaded speech must control the mouth movement. The input and evaluation criteria are clearer for that job.
No generative animation guarantees pixel-exact preservation. Treat the source image as a reference and inspect the full video for drift, substitutions, and new details.
Commercial use depends on your plan and your rights to the source image, people, brands, music, and other assets. Magic Hour’s paid plans permit commercial use under the current terms; free outputs are for personal, non-commercial use.
