Canva Craft, developed by Canva’s research team, generates images in two stages: it first interprets the request, then creates a plan defining spatial relationships, scale, visual hierarchy and embedded text before producing the image. The model complements the Canva Design Model, which creates layered, editable outputs and powers features such as Magic Layers and Canva AI 2.0. Canva also integrated Magic Layers with Google Gemini and ChatGPT in June 2026, previewed image-generation endpoints for its Connect API on October 2, 2026, and acquired Australian analytics startup MagicBrief on October 22, 2026. Before the acquisition, the platform reported 240 million monthly users and approximately $3 billion in annual revenue. By owning its model, Canva aims to control quality, costs and the integration of image generation with its editing tools while reducing reliance on model providers that are also becoming competitors.
Source: Read the original article
