Create
AI video from multiple images
Most image-to-video models animate one photo. H3 Max Reference takes up to 4 reference images (for example a person, a second person, a product and a location) and builds one new clip in which they appear together. It is the quickest way to put your own character into a scene with your own product.
Clips are 5 or 10 seconds at 480p, 768p or 1080p in 16:9, 9:16 or 1:1, without sound. From 25 credits (5s, 480p); 10 seconds at 1080p is 160 credits.
20 free credits on sign-up · failed jobs return credits automatically
Models that power it
| Model | Best for | Options | From |
|---|---|---|---|
| H3 Max Reference MiniMax | Up to 4 reference images (characters, products, places) in one video | 5/10s · 480p, 768p, 1080p | 25 cr 5s · 480p · ≈$0.50 |
| H3 Max MiniMax | MiniMax's newest video model, from 480p tests to 1080p | 5/10s · 480p, 768p, 1080p · start + end frames | 25 cr 5s · 480p · ≈$0.50 |
| GPT Image 2 OpenAI | OpenAI's sharpest images and typography, edits from up to 8 references | Medium, High · slow · up to 8 reference images | 8 cr Medium · ≈$0.16 |
| Nano Banana 2.1 Google | Google's newest image model, edits and up to 4K | 1K, 2K, 4K · up to 8 reference images | 5 cr 1K · ≈$0.10 |
Prices in credits for the cheapest settings. ≈$ at the Pro rate ($7.99 for 400 credits). Studios show the exact price before you generate.
How to do it, step by step
- 1
Open Video Studio with H3 Max Reference
The button on this page opens the studio with H3 Max Reference already selected; the credit price updates on the Generate button as you change settings.
- 2
Write one shot or add a start frame
Subject, one action, camera move, light. Upload 2–4 images, then refer to them by number: "the man from image 1 holds the bottle from image 2".
- 3
Test cheap, then finish
Run the shortest, lowest-resolution version first. Re-render only the take you like at a higher setting.
- 4
Download the MP4
Results land in Assets. Failed jobs return their credits automatically.
Example prompts
Starting points written for this task, not outputs we are showing you. Open one in the studio with its model selected, then adapt it.
The woman from image 1 walks into the bakery from image 3 and picks up the cake box from image 2, smiles at camera, 9:16
H3 Max Reference · from 25 crUse this prompt →The two friends from image 1 and image 2 clink coffee cups at a sunny window table, handheld, natural light
H3 Max Reference · from 25 crUse this prompt →The robot toy from image 1 rolls across the desk from image 2, camera tracks alongside at table height
H3 Max Reference · from 25 crUse this prompt →
What people make with it
Character + product
Your presenter holding your product.
Two people
Two portraits in one interaction.
Place a subject
Put a person into a specific location photo.
Storyboards
Quick moving drafts from concept stills.
FAQ
How many images can I combine?
Up to 4 reference images in one generation.
Will faces stay identical?
Reference video keeps strong resemblance but is not a guaranteed copy; check faces, logos and hands before publishing. For an exact product or face, image-to-video with a start frame is closer.
Does it have sound?
No. Add a voiceover, music or sound effects in the Audio studio afterwards.
Can I use photos of other people?
Only with their consent. Using a real person's likeness without permission is against the VYRA content policy.
How much does a multi-image video cost on VYRA?
From 25 credits per clip (about $0.50 on Pro, $7.99/month for 400 credits). The exact price for your settings is shown on the Generate button before anything is charged.
What if a generation fails?
The credits return to your balance automatically. You are only charged for finished results.
Try AI video from multiple images now.
Every model on one plan. Start with 20 free credits, upgrade when you need more.
Explore VYRA
AI tools
Models