Canva Magic Layers is Canva's latest AI feature, and it aims to turn flat images into editable designs that feel as if they were created natively inside the platform.
The tool takes a single-page PNG or JPG and uses artificial intelligence to break it into separate layers, including text, objects and background elements. Once processed, users can move, resize or delete individual components, change fonts, and adjust layouts without rebuilding the design from scratch. Canva says Magic Layers is designed to unlock AI-generated visuals that until now have largely been trapped in static files.
Magic Layers runs on the Canva Design Model, the company's proprietary foundation model created specifically for visual content. Introduced in October 2025, this model has already produced hundreds of millions of editable presentations, documents and social posts and also powers Canva's integrations with services such as ChatGPT, Claude and Microsoft Copilot. Instead of simply tracing outlines, the system interprets the overall structure of a design, restores text as live boxes and preserves relationships between elements so the layout remains coherent.
Canva Magic Layers is currently rolling out in public beta within the Canva editor for users in Australia, Canada, the UK and the US. During the beta, it supports single-page PNG and JPG files, with wider format support and geographic expansion planned in future updates. Canva positions the feature as part of its broader AI suite, targeting marketers, small businesses and creators who want faster, more flexible control over AI visuals rather than repeatedly prompting and waiting for new outputs.
For designers used to starting over each time an AI image missed the mark, Canva Magic Layers offers a more forgiving path: keep the idea, but edit the details. If the beta proves reliable across complex layouts and diverse languages, it is likely to become a routine tool for anyone relying on AI-generated graphics in everyday work.