Canva has released Canva Craft, its first in-house AI image generation model, developed by its research team to compete with similar tools from Google, OpenAI, ByteDance and Meta. According to TechCrunch and The Verge, the model is available now to all users globally at no extra cost, accessible through the Dream Lab tab in Canva's left-hand menu.
The launch deepens Canva's push into artificial intelligence and puts the design platform in direct competition with OpenAI, Microsoft and Google in the fast-growing market for AI image tools. Canva says the model was trained on millions of human-designed templates from its own library rather than on photographs.
How Canva Craft works: two-stage architecture
Canva Craft uses a two-stage system. In the first stage, it interprets a user's prompt and builds a plan covering spatial relationships, relative scale, visual hierarchy and text generation. In the second stage, it produces an image at up to 2K resolution in more than 30 possible aspect ratios for different platforms.
According to Canva, this approach delivers more accurate object details and legible text, and allows precise edits without altering other parts of the image. The model can generate social media posts, marketing content, realistic scenes, digital art and environmental imagery. Because it was trained on Canva's template library — posters, social posts, presentations, infographics, menus and typography-rich layouts — generated images can be converted into layered designs using Canva Magic, the company's image-to-design transformer.
- Up to 2K resolution image generation
- More than 30 aspect ratios supported
