Introducing WaveSpeedAI WAN 2.2 Fun Control on WaveSpeedAI

Introducing Wan 2.2 Fun Control: Precision Video Generation with Multi-Modal Control

The world of AI video generation has just taken a significant leap forward. We’re excited to announce the availability of Wan 2.2 Fun Control on WaveSpeedAI—a powerful video-to-video generation model that puts unprecedented creative control directly in your hands.

Developed by Alibaba’s PAI team and built on the advanced Wan 2.2 architecture, this model introduces innovative Control Codes that let you direct motion, structure, and scene composition with remarkable precision. Whether you’re creating dynamic character animations, transforming existing footage, or bringing static images to life, Wan 2.2 Fun Control delivers cinematic-quality results with the control you’ve always wanted.

What is Wan 2.2 Fun Control?

Wan 2.2 Fun Control represents the next evolution in controllable video generation. Unlike traditional text-to-video models that leave motion largely to chance, this model integrates deep learning with multi-modal conditioning to give creators fine-grained control over every aspect of video synthesis.

The model leverages Alibaba’s proprietary Mixture of Experts (MoE) architecture, which intelligently divides expert models according to denoising timesteps. This technical innovation results in higher quality video content with stable motion, consistent identity preservation, and professional-grade output that meets the standards of VFX, advertising, and virtual production workflows.

Trained on a significantly expanded dataset—65.6% more images and 83.2% more videos compared to its predecessor—Wan 2.2 Fun Control achieves top performance among both open-source and closed-source models across multiple dimensions including motion quality, semantic understanding, and aesthetic appeal.

Key Features

Multi-Modal Control System

The heart of Wan 2.2 Fun Control lies in its versatile control input options:

• Canny (Edge Detection): Guide object shapes and structural outlines using edge or line-art references
• Depth Maps: Control spatial relationships and 3D composition for scenes with natural parallax effects
• OpenPose (Skeletal Tracking): Direct human poses and movements with precise skeletal keypoint detection
• MLSD (Geometric Lines): Maintain architectural consistency and structural integrity in scenes
• Trajectory Control: Define exact motion paths for objects and camera movements

Cinematic-Quality Output

Built on the Wan 2.2 foundation, this model incorporates meticulously curated aesthetic data with detailed labels for lighting, composition, contrast, and color tone. The result is truly cinematic video generation with customizable aesthetic preferences and professional-grade visual fidelity.

Extended Video Duration

Generate videos up to 120 seconds in length at 720p resolution—a substantial capability that enables longer narrative sequences, complete dance routines, and extended action scenes without the need for stitching multiple clips.

Bilingual Prompt Support

Accept prompts in both Chinese and English, providing flexible creative control for international teams and diverse workflows.

Commercial-Ready Licensing

Released under the Apache 2.0 license, Wan 2.2 Fun Control can be freely integrated into commercial products without legal ambiguity, giving businesses complete flexibility in deployment.

Real-World Use Cases

Character Animation and Motion Transfer

Transform static character designs into dynamic animations by combining a reference image with motion guidance. Use a control video of a person dancing or performing actions, pair it with your character design, and the model intelligently merges them into a cohesive, style-consistent animation. This workflow is perfect for game developers, animators, and content creators who need to bring characters to life quickly.

Dance and Action Sequences

The Fun-Control architecture excels at generating extended dance and fight scenes. Community users have praised its ability to produce up to 2-minute sequences with consistent character appearance and natural motion flow—tasks that would traditionally require extensive manual animation work.

Marketing and Advertising Content

Create dynamic product showcases, brand videos, and promotional content with precise control over camera movements and scene composition. The trajectory control feature enables smooth camera paths that highlight products from multiple angles while maintaining professional visual quality.

Virtual Production and VFX

Film and television professionals can use Wan 2.2 Fun Control to generate pre-visualization content, prototype complex shots, or create background elements for compositing. The depth map control ensures accurate spatial relationships that integrate seamlessly with live-action footage.

Social Media and Content Creation

Transform existing video content or static images into fresh, engaging clips optimized for different platforms. The model’s ability to preserve structure while re-styling scenes makes it ideal for repurposing content across channels while maintaining brand consistency.

Getting Started on WaveSpeedAI

Using Wan 2.2 Fun Control on WaveSpeedAI is straightforward and production-ready from day one:

1. Navigate to the Model: Visit wavespeed.ai/models/wavespeed-ai/wan-2.2/fun-control
2. Upload Your Inputs: Provide your reference image or video along with your chosen control map (Canny, Depth, OpenPose, MLSD, or Trajectory)
3. Craft Your Prompt: Describe your desired output in English or Chinese
4. Select Resolution: Choose between 480p ($0.04/second) or 720p ($0.08/second)
5. Generate: Let the model work its magic

WaveSpeedAI’s infrastructure ensures you get results without the typical friction of AI model deployment:

• No Cold Starts: Your requests begin processing immediately
• REST API Ready: Integrate directly into your production pipelines
• Predictable Pricing: Billing is capped at 600 seconds (10 minutes) to keep costs manageable
• Affordable Entry: Minimum charges start at just $0.20 for 480p or $0.40 for 720p

Tips for Best Results

• Match Aspect Ratios: Keep input images and target video dimensions consistent for optimal output
• Maintain Composition Consistency: Ensure your reference image’s pose and camera angle align with your control map
• Start Simple: Begin with one or two control types and add more as you refine your workflow
• Experiment with Control Combinations: Different projects may benefit from layering Depth with OpenPose or combining Canny with Trajectory control

Take Your Video Creation to the Next Level

Wan 2.2 Fun Control represents a fundamental shift in how creators approach AI video generation. By putting precise control tools in the hands of artists, animators, marketers, and filmmakers, it bridges the gap between creative vision and AI-generated output.

The combination of multi-modal control inputs, extended video duration, cinematic quality, and commercial-friendly licensing makes this model an essential addition to any creative workflow. And with WaveSpeedAI’s reliable infrastructure, affordable pricing, and instant availability, there’s never been a better time to explore what controllable video generation can do for your projects.

Ready to take control of your video generation? Try Wan 2.2 Fun Control on WaveSpeedAI today and discover how precise, professional-quality AI video is now just a few clicks away.