Max duration

720p

Max resolution

↔

Start + End frame

Why pick Wan 2.2

↔ Both endpoints locked

Wan 2.2 is one of only two models in the lineup that lock BOTH start AND end frames. Define exactly where a clip begins and where it lands — critical for storyboarded animation sequences.

🔄 Seamless loops & chains

End frame of clip A = start frame of clip B. Chain multiple Wan 2.2 generations into seamless motion sequences — the only way to build longer videos from 5-second building blocks with matching cuts.

🔓 Uncensored output

Trusted users get unrestricted generation — no content filters on the animation. Start and end frame control with uncensored output in the same model.

🖼️ Precise transitions

When your storyboard defines both the opening and closing pose, Wan 2.2 fills the motion between them. More predictable than single-frame models for defined narrative beats.

⚡ Solid motion quality

Alibaba's open-source Wan architecture delivers reliable motion. Good physics, consistent subject tracking — a tested foundation even if the resolution ceiling is lower than newer models.

🗂️ Storyboard workflows

If your production uses a storyboard with defined visual beats — specific starting and ending poses — Wan 2.2 is the model to reach for among uncensored options.

What is Wan 2.2?

Wan 2.2 is Alibaba's open-source video generation architecture — one of the foundational models behind the Wan family on ZenCreator. Its defining capability among uncensored video models is dual-endpoint control: you can lock the start frame (the image the clip begins from) AND the end frame (the image it transitions to). The model fills in all the motion between those two reference points.

This makes Wan 2.2 specifically useful for storyboard-based workflows. If you have two images and need a clip that moves from one to the other — a pose transition, a position change, a facial expression shift — Wan 2.2 is one of only two models in the lineup that supports this. The other is Kling 2.1, which is censored; Wan 2.2 provides the same dual-endpoint capability with unrestricted output for trusted users.

The trade-off is resolution and duration. Wan 2.2 tops out at 720p and 5 seconds — lower than the 1080p you get from Wan 2.5, 2.6, and the Seedance family. For social media drafts and storyboard testing, 720p is workable. For final publish-ready content, you'll want to use a higher-resolution model once you've locked the motion direction.

See Wan 2.2 in action

Three examples showing the start → end frame transition workflow.

Wan 2.2 vs other ZenCreator video models

Model	Duration	Resolution	Start frame	End frame	Audio
Wan 2.2	5s	720p	✓	✓	—
Kling 2.1	5-10s	Variable	✓	✓	—
Wan 2.6 + Audio	15s	1080p	—	—	✓
Seedance Pro	10s	1080p	—	—	—
Kling 2.6	30s	—	—	—	—

When should you NOT pick Wan 2.2?

You need 1080p — Wan 2.2 tops out at 720p. For publish-ready 1080p video, use Wan 2.6 + Audio or Seedance Pro.
You need more than 5 seconds — Wan 2.2 clips are capped at 5s. For longer generation, use Wan 2.6 + Audio (up to 15s) or Kling 2.6 (up to 30s).
You don't need end frame anchoring — if you only have one source image, single-frame models like Seedance Pro will give better quality at 1080p. Wan 2.2's advantage is the dual-endpoint; if you're not using both, pick a higher-resolution option.

How to get started

Upload start + end photos

Write a transition prompt

woman is touching her feet

Get the transition clip

Open Image-to-Video→

Bottom line

Wan 2.2 has one defining advantage: it's the only uncensored video model on ZenCreator with both start AND end frame control. If your workflow requires precise transitions between two known images — a storyboard beat, a pose-to-pose cut, a loopable sequence — this is the model. Accept the 720p/5s ceiling in exchange for endpoint precision. For higher resolution output without endpoint anchoring, use Seedance Pro or Wan 2.6 + Audio.

Available in

Image-to-Video

Upload a start frame, optionally an end frame, write a transition prompt, pick Wan 2.2, generate.

Try Image-to-Video→

Text-to-Video

Generate video directly from a text prompt — no source image needed. Pick Wan 2.2 from the model list.

Open Text-to-Video

Start frame control means the clip begins from your source image. End frame control means the clip also ends on a second image you supply. Most models only support start frame — the motion ends wherever the model decides. With both endpoints locked, you define exactly what the clip transitions between, which is essential for storyboard-accurate animation and loopable sequences.

Both models support dual-endpoint control, but Kling 2.1 has safety filters permanently on — it won't process unrestricted content. Wan 2.2 is fully uncensored for trusted users. If you need start+end frame control with unrestricted output, Wan 2.2 is the only option.

Yes — this is the primary use case for the end frame feature. Set the end frame of clip N as the start frame of clip N+1. The result is a seamless chain of 5-second segments that cut together cleanly. Combine in any video editor.

720p is acceptable for Instagram Stories, TikTok, and most short-form platforms — the final compression at those platforms often reduces visible quality differences anyway. For YouTube or premium deliverables that need crisp quality, use a 1080p model instead.

For trusted users on ZenCreator, yes. Wan 2.2 runs without content filters for users with trusted access. Contact support to request trusted access if you don't have it yet.

Sources

Alibaba Wan model family technical details: wan.video
ZenCreator Image-to-Video tool: zencreator.pro
ZenCreator AI Models internal review database, June 2026

Wan 2.2