Canva has launched Canva Craft, its first image generation model. The company says it is designed around how people build visual layouts, with a planning stage before image creation and tools for making targeted edits afterward.
How Canva Craft builds an image
The model works in two stages. First, it interprets a prompt and maps out details such as where elements sit in relation to each other, their relative scale and the visual hierarchy. It also plans information needed to generate text within the image.
In the second stage, Canva Craft creates an image at up to 2K resolution. Canva says users can choose from 30 different aspect ratios, giving them options for images intended for different platforms.
This staged process is meant to help the model handle details that can be difficult for image generators, especially the placement of objects and readable text. Canva also says users can make precise changes without altering other parts of an image. That could make it easier to refine a layout while keeping the elements that already work in place.
Training on design layouts
Canva says it trained Craft on design rather than photographs. Its training material included human-designed templates from the Canva library, among them posters, social posts, presentations, infographics, menus and layouts with substantial typography.
That choice connects image generation to the kinds of designed assets people already create on the platform. A generated image can be converted into a layered design using Canva Magic, the company’s image-to-design transformer. Canva released its own model for understanding designs and layers last year.
Layered designs give users a way to work with the parts of a visual separately. In practical terms, Canva’s approach links the act of generating an image with further editing inside a design tool, rather than treating the output only as a finished picture.
Where users can access it
Canva Craft is available to all users through the Dream Lab tab on Canva’s platform. Canva also plans to make the model available through its API, and says it intends to support image generation in third-party chatbots such as ChatGPT and Gemini through Canva’s MCP.
The company already uses a range of image models in its platform. It says Craft offers more control over quality and speed, in addition to potential cost benefits. Using its own model could help Canva manage the costs associated with image generation, though the company’s statement describes this as one of several benefits rather than a separate product feature.
Researchers from Leonardo.ai contributed to the model. Canva acquired the media generation platform in 2024, adding that team’s experience to its research effort.
More versions are planned
Canva says it is developing two versions designed to be more cost-efficient: Craftfast and Craft-ultrafast. The company says both will be released soon, but has not provided further details in the source announcement.
The launch comes as Canva faces a changing business outlook. The company was on track for $4 billion in revenue last year, but cut its growth forecast for 2026. External analysts reportedly placed its valuation at $35 billion, down from $42 billion.
Against that backdrop, Canva Craft serves two roles in the company’s platform: it adds an in-house option for generating images and connects those images to Canva’s existing design and editing tools. Its reach may extend beyond Canva’s own interface if API and chatbot support arrive as planned.