Canva launches in-house image generation model Craft to compete with major AI players

Canva released Canva Craft, its first in-house image generation model, trained by Canva Research primarily on millions of design templates rather than photos. Craft can generate images up to 2K resolution across over 30 aspect ratios, render text in 55 languages, and allow users to break results into editable layers using Magic Layers. On the DesignArena platform, Craft ranked fourth overall, and first in portraits and landscapes. Craft is currently available to all users via Dream Lab and the public API, with MCP integration for chatbots like ChatGPT and Gemini planned later. Researchers from Leonardo.ai, acquired by Canva in 2024, also contributed to the project, and cheaper variants Craft-fast and Craft-ultrafast are expected soon.