Canva has launched its first proprietary image generation model, Canva Craft, which distinguishes itself by planning a design before creating the final visual. This two-stage architecture aims to provide greater precision, consistency and control compared to standard generative tools. The system produces images at up to 2K resolution across more than 30 aspect ratios. It targets social media posts, marketing content, realistic scenes, digital art and environmental imagery.
The approach matters because it addresses common frustrations with current AI image generators, specifically the lack of structural logic in composition. By separating the planning phase from the rendering phase, the model attempts to reduce random errors and maintain visual coherence throughout a project. This method aligns closely with how professional designers actually work, using sketches or wireframes before finalising artwork.
* Two-stage process separates planning from generation
* Supports over 30 different aspect ratios
* Targets 2K resolution output
