Canva has introduced Canva Craft, a new artificial intelligence image generation model developed to compete in the market dominated by major technology companies. The model employs a two-stage architecture that distinguishes it from competing solutions by planning designs before generating images, a feature the company says enables greater precision, consistency, and user control over the final output.
The model supports image generation at resolutions up to 2K and offers more than 30 different aspect ratios for various use cases. Beyond social media content and marketing materials, Canva Craft is designed to create realistic scenes, digital artwork, and environmental imagery. The two-stage system allows for more accurate generation of complex images, particularly those requiring legible text and precise object placement.
Canva developed Craft using training data drawn from millions of human-made templates stored in its content library, which encompasses social media posts, presentations, infographics, and menu designs. According to the company’s announcement on Thursday, the model currently ranks fourth on Design Arena’s leaderboard for AI image generation models. The model incorporates built-in safety features including input and output moderation, privacy controls, and reporting capabilities.
The launch continues Canva’s broader artificial intelligence expansion. Earlier in April, the platform introduced prompt-based editing tools and added layering capabilities to AI-generated images, enabling users to modify specific portions of designs without affecting other elements. Looking forward, Canva plans to enable AI agents to integrate with Craft through the Model Context Protocol and will release optimized versions called Craft-fast and Craft-ultrafast.
Canva Craft is now accessible globally to all platform users through the Dream Lab feature, which can be accessed via the three-dot menu on the left-hand sidebar.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI