Photo-to-video is the core of what Rigel does. You give the app one image; the app gives you back a short film of it. This page explains what actually happens, what you can realistically get out of it, and where the limits are.
How photo to video works in Rigel
Three steps, in the app, with nothing installed on a computer.
- Upload a photo. A Live Photo, a portrait, a screenshot of an illustration — any still image in your library.
- Describe the scene. One sentence is enough: who or what is in frame, what they do, where they are, how it is lit. If you want the subject to speak, paste the script here too.
- Generate and share. Rigel renders a 5–30 second clip, vertical or square, ready to post.
The prompt matters more than people expect. "Woman smiles" is generic. "Slow push-in on a woman turning to camera and smiling, late afternoon window light, shallow depth of field" gives the model camera movement, timing and lens behaviour to work with.
What you can make from one still
- Animated portraits — subtle head turns, blinks, breath and hair movement that keep the person recognisably themselves.
- Talking videos — add a script and get realistic lip-sync, so the subject appears to speak your words. See the talking photo guide.
- Expression changes — smile, react, tear up, perform. Useful for character work and for reaction-style social clips.
- Scene generation — place the subject on a beach at sunset, a rainy street, a dark studio, wherever the prompt says.
- Multi-scene stories — switch location or shot mid-video for story-driven Reels and Shorts. See long AI video.
The models behind the output
Rigel is a product built on top of frontier video models, not a model lab. That distinction is deliberate: the generation pipeline gets updated as better models land, so the quality ceiling of the app moves with the field rather than with our release cycle.
Two model generations reset expectations recently — Google's Veo 3 and Seedance 2.0. Both pushed temporal consistency, physics-plausible motion and prompt fidelity forward, which is exactly what photo-to-video depends on. We wrote up what each one changed on the Veo 3 page and the Seedance 2.0 page.
Where it works well, and where it doesn't
It works well on single clear subjects, portraits, pets, product shots and illustrations. It struggles — as every model in this class currently does — with dense crowds, text inside the image, complex hand interactions and very long uninterrupted takes. When a result misses, the fastest fix is usually a more specific prompt or a cleaner source photo, not a regeneration of the same inputs.
What it costs
The app is free to download and comes with starter credits, so you can see your own photo animated before paying anything. After that, credit packs and a subscription are available in-app. Every generation shows its credit cost before you confirm, and credits from packs do not expire.
Rigel is rated 4.6★ on the App Store and 4.7★ on Google Play across 600+ reviews, and runs on iOS 16.6+ and Android. Download Rigel and start with a photo you actually care about — the source image matters more than any prompt trick.