Key facts
- Canva Craft interprets prompts, plans image composition, and generates images up to 2K resolution.
- The model was trained on designs from Canva's library, not photographs.
- Canva Craft can be integrated into third-party chatbots.
- Canva acquired Leonardo.ai in 2024, contributing to the new model's development.
Canva has launched Canva Craft, its inaugural AI image generation model, developed internally to rival offerings from major tech companies like Google, OpenAI, ByteDance, and Meta. The model operates in two stages: first, it interprets user prompts to plan image elements, and second, it generates images up to 2K resolution in various aspect ratios. This approach allows for detailed objects and legible text, with precise editing capabilities. Unlike models trained on photographs, Canva Craft was trained on designs from Canva's extensive library, enabling generated images to be converted into layered designs via Canva Magic. The model is accessible through the Dream Lab tab on Canva's platform and via its API, with plans for integration into third-party chatbots such as ChatGPT and Gemini. The company also stated that the new model offers better quality and speed control, potentially reducing API costs. Canva acquired Leonardo.ai in 2024, and its researchers contributed to the development of Canva Craft. The company is also developing faster, more cost-efficient versions, Craftfast and Craft-ultrafast.