Canva’s AI image story developed across tools rather than one single model. Its 2023 Magic Edit, Eraser and Expand capabilities target existing designs; its October 2024 Dream Lab introduced another route for generating image concepts.
Use the right operation
Canva’s 2023 retrospective identifies Magic Edit as natural-language image modification, Magic Eraser as object removal, and Magic Expand as extending an image beyond its frame. In a marketing workflow, an editor can swap a background prop without rebuilding the entire poster. That operation is different from creating a wholly new image and does not guarantee that the altered output preserves every material detail of the original.
Dream Lab and Leonardo context
Canva’s 2024 retrospective dates its acquisition of Leonardo.Ai to July and the Dream Lab launch to October. The retrospective describes the integration of Leonardo technology and the Phoenix model into Canva’s AI suite. The existence of a foundation-model relationship does not prove the exact model snapshot behind each Canva image feature or that every Leonardo control is exposed to Canva users.
Team production and permissions
Use Canva to assemble layouts with actual text, approved font choices and licensed stock assets. Generate concept art where useful, then review logos, face likeness, false endorsements and claims implied by photorealistic scenes. A brand kit is a visual consistency aid, not a legal trademark search or a guarantee of identical generated imagery. Check collaboration permissions before uploading unreleased products or customer portraits.
Changing plans and regions
The dated source material establishes historical capabilities but does not by itself establish October 2026 tier limits, market eligibility or credit pricing. Avoid generic “free forever” or unlimited language. Record the viewer-facing Canva app, region, subscription tier and available AI credits against AI 10’s pricing ledger, and attach screenshots only if your team has permission and actually captured them.
Worked scenario and failure test
A restaurant chain wants one campaign key visual for menus, banners and a social story. Use a design template with approved actual copy, and create decorative imagery through Dream Lab or a suitable licensed source. Use Magic Edit only for changes to parts of an already approved picture, such as an unwanted prop. If a generated dish looks unlike the item sold, replace it with the real photographed dish. The brand reviewer should sign off on product truthfulness and allergy or price text. Current Canva features cannot be assigned a specific Leonardo model ID without additional product-level evidence.
Decision guide and practical handoff
Use generative tools for imagery and conventional Canva layers for final claims and text. A responsible review packet names the exact source or model version, the permitted input assets, the requested operation and the image characteristics a reviewer must check. It should also contain an unaltered reference, the accepted derivative, and a short record of what changed. A supported feature is not necessarily a reliable one for every brief, and the team must approve both technical appearance and real-world truthfulness before distribution.
Availability, price and geographic limitations
Research cutoff: 25 October 2026. Release dates and historical tier details above are explicitly dated. Unless independently confirmed for a named country and account tier, current availability, billing currency, credit consumption, API quota and commercial entitlement are unknown—not universal. Consult official regional terms and AI 10’s dated price snapshots before procurement; do not treat a historical announcement as a live entitlement. No application, model, API response or image output was hands-on tested for this article.
Sources & Further Reading
- Canva image AI 2023 retrospective — published/updated 2023 (consulted at research cutoff 2026-10-25).
- Canva 2024 / Dream Lab retrospective — published/updated 2024 (consulted at research cutoff 2026-10-25).