How to turn a photo into a video with Gemini Omni
Gemini Omni can turn a single photo — or up to seven — into a short, moving video. It's one of the easiest ways to bring a portrait, product shot, or piece of art to life. Here's how to do it on Dora.
Step 1 — Open Gemini Omni
Go to the create page and select Gemini Omni from the model menu. It's an "any-to-any" model, so the same entry handles text, photos, and video.
Step 2 — Add your photo(s)
Upload up to 7 reference images. More references help when you want to:
- keep a person or character consistent,
- lock a product so it doesn't get redesigned, or
- establish a visual style (lighting, palette, mood).
For a single subject, one clean, well-lit photo is usually enough.
Step 3 — Write a motion prompt
The photo sets the look; the prompt sets the motion and camera. Be specific about what should move:
"Slow push-in on the subject, hair moving gently in the wind, soft golden-hour light, shallow depth of field."
Avoid contradicting the photo (don't ask for a different outfit or background unless you want it changed).
Step 4 — Pick resolution, duration and aspect ratio
- Resolution: 720p for speed and drafts, 1080p or 4K for final output.
- Duration: 4–10 seconds. Shorter clips tend to look cleaner.
- Aspect ratio: 9:16 for Reels/TikTok, 16:9 for YouTube/landscape.
Step 5 — Generate and iterate
Generate, review, and tweak the prompt. Small changes ("add a subtle camera orbit", "more dramatic lighting") often improve the result more than rewriting everything.
Tips for better photo-to-video
- Match the aspect ratio to your photo to avoid awkward cropping.
- Keep motion realistic for portraits — gentle is more believable than fast.
- Use multiple angles of the same subject for stronger consistency.
- For faces that need to speak, use Dora's Lip Sync model instead.
Related models on Dora
If Gemini Omni isn't quite the look you want, SeeDance 2 is excellent for cinematic image-to-video, and Kling 2.6 does image-to-video with audio.