Canva launches its 1st image generation model
What's the story
Canva, the popular design platform, has launched its first-ever image generation model called "Canva Craft." The innovative tool was developed by the company's research team to compete with similar offerings from tech giants like Google, ByteDance, OpenAI, and Meta. The new model works in two stages: interpreting user prompts and generating high-resolution images across different aspect ratios.
Training method
How Canva Craft is different from other models
Unlike other models that rely on photographs, Canva Craft has been trained using human-designed templates from the company's library.
These include posters, social posts, presentations, infographics, menus, and typography-rich layouts.
This unique training method ensures that the images generated by Canva Craft are detailed and text is legible.
Users can also make precise edits without affecting other parts of the image.
Integration features
It can be accessed via the Dream Lab tab
Canva Craft is accessible to all users on the company's platform via the Dream Lab tab. It will also be available through the Canva API.
The company plans to integrate image generation into third-party chatbots such as ChatGPT and Gemini via its MCP.
This integration could potentially reduce API costs for Canva while improving quality and speed control on its platform.
Acquisition impact
Co-developed with Leonardo.ai researchers
In 2024, Canva acquired media generation platform Leonardo.ai.
The company has revealed that researchers from Leonardo contributed to the development of this new image model.
Looking ahead, Canva is working on two cost-efficient versions of the model, dubbed Craftfast and Craft-ultrafast, which are expected to be released soon.