Wan 2.2 Uncensored
Wan 2.2 Uncensored is Alibaba Tongyi Wan's video model, hosted on Vellria with nothing to install: text to video and image to video, 480p to 720p, clips of 5–8s. Run it in Studio or over the REST API, from 8 credits / s (100 credits = $1.00).
vellria-wan-2-2Wan 2.2 is developed by Alibaba Tongyi Wan. Vellria provides access to the model; it does not own it. All trademarks belong to their respective owners.
Output
No generation yetCover art
from 8 credits / s
Your request will cost 40 credits.
About Wan 2.2
Also known as: Wan 2.2 NSFW · Wan2.2 · Alibaba Wan 2.2 · Wan 2.1 successor
Wan 2.2 is the open-weight video release from Alibaba's Tongyi Wan team, and most local NSFW video workflows are still built on it: the checkpoint people load in ComfyUI, pair with LoRAs and leave running overnight on a large graphics card. Here it runs hosted, from a written prompt or from a picture, with nothing to download and no VRAM to budget.
The clips are deliberately simple. You pick five or eight seconds, 480p or 720p, and for written prompts a landscape or portrait frame. There is no soundtrack, which keeps the price per second among the lowest in the catalog, so Wan 2.2 is the inexpensive place to try many variations of a scene: the wording, the camera angle, the pacing.
Text-to-video and image-to-video ask for different prompts. A written prompt invents everything, so name the subject, the setting, the single action and the camera move. With an uploaded still, the picture has already decided who and where; spend the words on motion instead: what moves, how fast, and where the camera travels.
Wan 2.2 is not the newest member of the family. Wan 3.0 is the current generation, with sound and much longer single takes, and Wan 3.0 Pro is its finishing tier. A practical pattern is to find the motion cheaply here and rerun the idea on Wan 3.0 when it needs audio or more time. For stills from the same family, Wan 2.7 Image makes opening frames worth animating.
Wan 2.2 in action
Wan 2.2 outputs with the exact prompts behind them.
Image to video
Her braided hair and the flowing turquoise gown sway gently in the current as silver fish swirl around her and she slowly raises her gaze to the camera, shafts of sunlight shimmering, slow push-in
Image to video
She glides forward on her roller skates past the palm trees, her hair bouncing in the breeze, the pink sunset glowing behind her, smooth tracking shot
Text to video
A woman in her thirties in an emerald satin evening gown walks slowly through a candlelit art deco hotel lobby and glances over her shoulder at the camera with a soft smile, warm golden light, smooth tracking shot
When to pick Wan 2.2
Pick it when
- You know Wan 2.2 from local ComfyUI workflows and want the same model hosted, without downloads or a graphics card.
- You want many inexpensive test clips to settle wording, motion and camera angle before spending more.
- A short silent clip of five or eight seconds at 480p or 720p is all the scene needs.
Pick another model when
- The clip needs sound or a single take far longer than eight seconds; use Wan 3.0.
- The shot is decided and should look its best at a higher resolution; finish it on Wan 3.0 Pro.
- You want quick drafts that come with generated sound; try LTX 2.5 Fast.
Prompt or reference
Write the scene from scratch to invent the shot and choose landscape or portrait. Start from a still to keep a look you already have: it becomes the opening frame and the clip follows its proportions, so crop before you upload.
How to use Wan 2.2 on Vellria
- Prompt: up to 10,000 characters in the panel above.
- Cost: from 8 credits / s — see pricing.
Example prompts
An adult woman with wet hair steps out of a steaming outdoor bath at dusk, slow camera pan following her, paper lanterns glowing behind.
Portrait frame: an adult man in a dim hotel room loosens his tie and sits on the edge of the bed, static camera, warm lamp light.
From the uploaded still: she turns her head toward the camera and smiles, hair lifting in a light breeze, gentle push-in.
Two adult lovers kiss slowly on a balcony at night, city lights blurred behind them, camera circling at shoulder height.
An adult dancer in a sheer robe spins once under a single spotlight on an empty stage, wide static shot, fabric trailing behind her.
Tips
- Write one action per clip; five or eight seconds is room for a single movement, not a sequence of events.
- Explore wording and camera moves at 480p, then rerun the version you like at 720p.
- Leave out LoRA trigger words from local workflows and describe the look in plain language.
- Name the camera move outright, such as a slow push-in, a pan or a locked-off frame, or the model chooses one for you.
- Describing sound changes nothing here; if the scene needs audio, take the idea to Wan 3.0.
Specifications
| Resolution: | 480p720p |
|---|---|
| Duration: | 5–8s |
| Aspect ratio (Text to video): | 16:99:16 |
| Reference images: | 1 reference image |
See also: uncensored AI video generator · NSFW video generation · NSFW image to video.
Pricing
| Resolution | Credits |
|---|---|
| 480p | 8 credits / s |
| 720p | 15 credits / s |
Wan 2.2 vs Wan 3.0 and Wan 3.0 Pro
Wan 2.2 is the open-weight Wan release that many people run locally, hosted here without downloads. It is the silent, lowest-priced member of the Wan video line on Vellria; Wan 3.0 and Wan 3.0 Pro add sound and much longer takes.
| Model | Price | Max resolution | Duration | Reference images |
|---|---|---|---|---|
| Wan 2.2 Uncensored This page | from 8 credits / s | 720p | 5–8s | Up to 1 |
| Wan 3.0 Uncensored | from 15 credits / s | 1080p | 2–30s | Up to 1 |
| Wan 3.0 Pro Uncensored | from 20 credits / s | 1080p | 2–30s | Up to 1 |
Prices and limits come from the live model catalog.
Wan 2.2 or Wan 3.0?
- Price: from 15 credits / s (Wan 2.2: from 8 credits / s)
- Max resolution: 1080p (Wan 2.2: 720p)
- Duration: 2–30s (Wan 2.2: 5–8s)
- Prompt length: 7,500 characters (Wan 2.2: 10,000 characters)
- Settings: Audio (Wan 2.2: None)
Wan 3.0 is the step up when a clip needs sound or more time: the same lineage with a higher resolution step, longer takes and a higher price per second. Test the motion idea here, then rewrite it as one continuous description for Wan 3.0.
Wan 2.2 or Wan 3.0 Pro?
- Price: from 20 credits / s (Wan 2.2: from 8 credits / s)
- Max resolution: 1080p (Wan 2.2: 720p)
- Duration: 2–30s (Wan 2.2: 5–8s)
- Prompt length: 7,500 characters (Wan 2.2: 10,000 characters)
- Settings: Audio (Wan 2.2: None)
Wan 3.0 Pro is the finishing tier of the newer generation. Go there directly only when the shot is already decided; otherwise Wan 3.0 is the cheaper middle step.
Wan 2.2 API: price and a first request
Wan 2.2 costs from 8 credits / s over the API, the same rate as in Studio. Model id vellria-wan-2-2 on POST /v1/generate/video.
curl https://vellria.com/v1/generate/video \
-H "Authorization: Bearer vll_..." \
-H "Content-Type: application/json" \
-d '{
"model": "vellria-wan-2-2",
"prompt": "…",
"duration": 5,
"resolution": "480p",
"ratio": "16:9"
}'Parameters, polling and uploads: API documentation · NSFW video API.
FAQ
Does Wan 2.2 allow NSFW?
Yes, for fictional adult content. Nudity and explicit scenes between adults are allowed. Minors are refused automatically and real people may not be animated without consent.
Is Wan 2.2 free?
No, there is no free plan, but a new account created with Google gets 20 credits as a gift, valid for 7 days, which you can put toward Wan 2.2 at from 8 credits / s. Beyond that every run is paid from prepaid credits at the rate this page shows before you start, and a run that fails at the provider is refunded automatically. Is there a free uncensored AI generator? covers the cheapest way to a first result.
Is this the actual Wan 2.2 model?
Yes. It is Wan 2.2 from Alibaba's Tongyi Wan team, run by our providers rather than on your own machine. Both written prompts and uploaded pictures are supported.
Why only five or eight seconds?
Those are the two lengths Wan 2.2 is offered at, for written prompts and uploaded pictures alike. For longer single takes with sound, use Wan 3.0.
Can I use my Wan 2.2 LoRAs?
Not at the moment. The hosted model runs without custom LoRAs, so describe the style, the body and the setting in words; a precise prompt gets close to most looks people reach for with a LoRA.
Should I use Wan 2.2 or Wan 3.0?
Use Wan 2.2 for cheap, short, silent drafts and Wan 3.0 when the clip needs sound, a longer take or a higher resolution.
Can I generate Wan 2.2 clips by API?
Yes. The API section on this page shows the request; see the API reference for polling and downloads.
Related models
Make the start frame for your clip with one of our image models.
Jib Mix Qwen Image Uncensored
Uncensored 1K portraits with natural, attractive faces from a Qwen-Image fine-tune, in seven shapes at one price.
Image
5 credits
Z-Image Spicy Uncensored
Uncensored Z-Image pictures from text in seven shapes up to 1536 px, all at one flat low price.
Image
4 credits
Z-Image Spicy Pro Uncensored
Larger uncensored Z-Image pictures from text, up to 2560 px in seven shapes at one flat price.
Image
5 credits