Canva’s first AI image model plans designs before generating them
The Verge Emma Roth ● Covered by 2 sources
Canva launched its first image model, Canva Craft. It plans a design before making the image, aiming for more control.
Based on reporting by The Verge, Emma Roth — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Canva is pushing further into AI with a new image model called Canva Craft, and it is aiming straight at tools from OpenAI, Microsoft, and Google. The big idea is simple enough: the model does not just spit out an image, it plans the design first, then generates it.
That two-stage setup is supposed to give users more precision, consistency, and control. Canva says the model can produce images at up to 2K resolution and across more than 30 aspect ratios, which matters if you are trying to make something fit a social post, a marketing asset, or a different format without starting over.
The pitch also stretches beyond the usual ad-style output. Canva says Craft is meant to handle realistic scenes, digital art, environmental imagery, and other kinds of visuals, alongside social media posts and marketing content.
This is Canva doing what a lot of software companies are now doing: taking a feature users already want and wrapping a lot more AI around it. The difference here is that Canva is leaning on planning, not just generation, as the selling point. That is a more useful promise than most AI copy, which is a low bar, but still.
My take — AI-written commentary, not fact-checked reporting
Canva is making the sensible bet here: people want fewer random image outputs and more control. The industry loves calling every new model magic; users usually just want the thing to land in the right aspect ratio without a small emotional crisis.
Read more about this at: The Verge
