Wan22.io: Open-Source Cinematic Video Generation

Wan22.io is an innovative open-source platform from Alibaba Tongyi Lab designed to empower creators and businesses with cinematic AI video capabilities. It features the world's first open-source Mixture-of-Experts (MoE) architecture for video diffusion models, enabling seamless text-to-video and image-to-video generation at 720P resolution with professional audio integration and precise visual control.

Key Features:

• Open-Source MoE Architecture: World's first open-source Mixture-of-Experts model for cinematic video diffusion.
• Text & Image to Video: Generate videos from text (T2V-A14B) or images (I2V-A14B) at 720P/24fps resolution.
• GPU Optimization: Runs efficiently on consumer-grade GPUs like RTX 4090 via TI2V-5B model.
• Cinematic Vision Control: Advanced lighting, color grading, and composition adjustments for professional results.
• Professional Audio Generation: Creates immersive soundscapes integrated with video output.
• User Content Ownership: Retain full rights to all generated content; underlying AI remains the platform's property.
• GitHub Accessibility: Complete model weights available openly on GitHub for community use.

Pricing: Wan22.io operates a freemium model where core open-source capabilities are free, but their hosted web service offers three subscription tiers: Basic ($20.9/month or $251/year, 9,800 credits), Plus ($34.9/month or $419/year, 25,200 credits), and Enterprise ($69.9/month or $839/year, 59,400 credits). All plans include no watermarking and commercial licensing rights.

Conclusion: Wan22.io redefines accessible AI video creation by combining open-source innovation with professional-grade tools, delivering cinematic quality without compromising user ownership or creative freedom.