Canva Launches Craft AI Model for Design-Centric Image Generation
Canva has released Canva Craft, its first proprietary image generation model, designed to compete with offerings from major tech firms. Unlike competitors, Craft is trained on human-designed templates rather than photographs, enabling precise text rendering and layered design conversion. The tool supports 2K resolution images across 30 aspect ratios and is available to all users via the platform and API, marking a strategic shift toward design-specific AI capabilities.
Key points
- Canva, a global graphic design platform, released Canva Craft, its first in-house image generation model, to compete with AI tools from Google, OpenAI, ByteDance, and Meta.
- The model operates in two stages: first interpreting prompts to plan spatial relationships and visual hierarchy, then generating images up to 2K resolution in 30 different aspect ratios.
- Unlike competitors that train on photographs, Canva Craft was trained on human-designed templates from Canva's library, including posters, infographics, and typography-rich layouts, to ensure legible text and accurate object details.
- Images generated by Craft can be converted into editable, layered designs using Canva Magic, a feature released last year that understands design structures.
- Canva Craft is currently available to all users through the Dream Lab tab and via the Canva API, with plans to integrate the model into third-party chatbots like ChatGPT.
- The release positions Canva to offer precise, design-oriented AI tools that allow users to make specific edits without altering other parts of the image.
Canva has launched Canva Craft, its first proprietary image generation model, developed by its internal research team to compete with established AI image generators from Google, OpenAI, ByteDance, and Meta. The company stated that the model is designed to address specific limitations in current AI image tools, particularly regarding text legibility and the structural integrity of design elements. By focusing on design-specific outputs, Canva aims to differentiate its offering from general-purpose image generators that often struggle with complex layouts.
The model functions in two distinct stages. Initially, it interprets user prompts to create a detailed plan, defining spatial relationships, relative scales, and visual hierarchies. In the second stage, it generates high-resolution images up to 2K across 30 different aspect ratios suitable for various digital platforms. Canva noted that this approach allows for precise edits to specific elements without disrupting the rest of the image, a capability critical for professional design workflows.
A key distinction for Canva Craft is its training data. Rather than relying on photographs, the model was trained on human-designed templates from Canva’s extensive library, including posters, social media posts, presentations, and infographics. This foundation enables the generation of images with accurate object details and legible text. Furthermore, images created with Craft can be converted into editable, layered designs using Canva Magic, the company’s image-to-design transformer released last year. Canva Craft is now available to all users via the Dream Lab tab and through the Canva API, with future integration into third-party chatbots planned.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.
