Turn one sentence into a Wan 2.2 clip
Describe a scene in plain words and the model brings it to life as a short clip. You need no photo and no footage to start. Test an idea in minutes, then sharpen the prompt and run it again.
Wan 2.2 AI video generator online
Type a detailed prompt or add one start photo, and Alibaba's Wan 2.2 returns a clip in 16:9, 9:16 or 1:1 in about a minute and a half.
Model
Wan 2.2 on Pixelcut
Wan 2.2 by Alibaba runs here as its A14B text to video and image to video models. Write a detailed prompt, add a start photo if you have one, and get a five second MP4.
Wan 2.2 at a glance
- Input
- Text or photosOptional start photo
- Resolution
- 1280 × 1280
- Aspect ratio
- 1:1, 9:16, and 16:9Landscape, vertical or square
- Output
- MP4
- Clip length
- 5 s
- Audio
- No sound
Browse Wan 2.2 video ideas
Each tile is a short Wan 2.2 clip. Take its subject, setting or light into a prompt of your own.
Prompt or photo
Choose Wan 2.2 text to video or a start photo
Start from words when the scene only exists in your head and you have nothing to film. Start from a photo when you already have the shot you want to move. Both paths give you the same kind of MP4, ready to post or drop into an edit.
- Pick text to video or image to video
- Add a photo only if you have the opening shot
- Describe what happens, then generate
A red ladybird on a green fern opens its spotted shell, unfolds its wings and lifts off, slow push in
Set the first frame with Wan 2.2 image to video
Upload one photo and your clip opens on that picture. You start from a shot you already like instead of a guess. Your prompt only has to say what moves, and the model sets it in motion.
Shape Wan 2.2 AI clips for phones, feeds and YouTube
Make it vertical 9:16, square 1:1 or wide 16:9 before you generate. Your clip fills a phone screen for Reels or plays wide on YouTube. It comes out in the shape you need, with no cropping after.
Write a detailed Wan 2.2 prompt for the shot you pictured
Describe the whole shot in up to 2,000 characters on Pixelcut. Put the subject and its action first, then the place and the light. You get room to plan every detail of the scene before you generate.
Use Wan 2.2 online for your next project
Cutaways for YouTube editors, creature pitches for indie game makers and openers for wedding and festival invites.
YouTube video editors
Fill a gap in your edit with a shot you never filmed. Make a wide cutaway that sits inside a long video. Find more video models in the AI video generator.
Indie game makers
Pitch a boss or a creature before you build it. Get a moving shot with a trailer feel to show your team or your backers.
Wedding and festival invites
Open a digital invite or a WhatsApp status with a moving reveal. Your guests get a warm welcome that feels made for the big day.
More models on Pixelcut
Switch models per generation instead of per subscription. Each card opens that model's page.
Cinematic motion, and your first video is free
- New
- Free with sign-up
- Image to video
- Up to 10 s
ByteDance's newest video model
- New
- Audio
- Image to video
- Up to 30 s
Text, image and reference video, up to 15 s
- New
- Audio
- Image to video
- Up to 15 s
Photoreal video with synchronized audio
- Audio
- Image to video
- Up to 8 s
Multi-shot scenes and elements
- New
- Audio
- Image to video
- Up to 15 s
Synchronized audio in many languages
- Image to video
- Up to 10 s
How to use the Wan 2.2 AI video generator
- 1
Write what should happen
Put the subject and its action in the prompt box in plain words. Have a picture that sets the scene? Add it as the opening frame and describe only the motion.
- 2
Set the aspect ratio
Pick a tall, square or wide shape before you press generate. Match it to the screen most people will watch the clip on first, whether that is a phone or a laptop.
- 3
Generate and save the MP4
Press generate and save the MP4 to your device when it is ready. If the result is close but not quite right, edit one line of the prompt and run it again.
Plans for video generation
Video is billed in credits per second of footage. Pick the plan that matches how much you make.
Save 20% with a yearly plan
Free
$0
per month, billed yearlyYour first video free on MiniMax H3 Max
Image tools free to try
Pro
$10$8
per month, billed yearly600 AI Credits Monthly
Access all AI models
Unlimited Background Removal
Unlimited Upscale
3 Person Team
1,000 Batch Exports Monthly
Commercial license
Most Popular
Business
$30$24
per month, billed yearly3,600 AI Credits Monthly
Access all AI models
Unlimited Background Removal
Unlimited Upscale
10 Person Team
2,000 Batch Exports Monthly
Commercial license
Trusted by over 200,000 creators, brands and small businesses
Trusted by creatives at
Frequently Asked Questions
Wan 2.2 is a video model from the Wan family by Alibaba's Tongyi lab. Pixelcut runs its A14B text to video and image to video models, alongside Wan 2.5, a later Wan version.
Yes. Add one start photo to your prompt and the model opens the clip on that picture, then moves it the way you describe. A prompt with no photo works as well.
Every clip on Pixelcut runs five seconds and comes as an MP4. The length is the same on every run. You pick the shape before you generate.
Yes. This model is part of Pixelcut Pro, and text to video and image to video both run on a Pro plan. The pricing section on this page lists the plans and what they include.
Yes, in most cases. Ads, product listings, social posts and client work are all fine uses, as long as you own the rights to the photo or words you put in and follow Pixelcut's policies.
Yes. The Pixelcut MCP server brings Wan 2.2 to Claude, ChatGPT, Codex and Cursor. Each request can be a prompt by itself or a prompt with one start photo, and clips use Pixelcut credits.
