Canva Releases Canva Craft, Its First In-House Image Generation Model

3.7 · 9d ago · WORLD · AI

The Facts

  • Canva released its first image generation model, called Canva Craft, developed by its research team. TechCrunch only
  • Canva Craft works in two stages: it first interprets the prompt and plans spatial relationships, relative scale, visual hierarchy and text information, then generates an image of up to 2K resolution in one of 30 aspect ratios. TechCrunch only
  • The model was trained on human-designed templates from Canva's library (posters, social posts, presentations, infographics, menus, typography-rich layouts) rather than photographs. TechCrunch only
  • Images generated with Canva Craft can be converted into layered designs using Canva Magic, the company's image-to-design transformer. TechCrunch only
  • Canva Craft is available to all users on Canva's platform via the Dream Lab tab and will also be offered through the Canva API. TechCrunch only
Dive deeper · 5 more factsShow less
  • Canva plans to support image generation in third-party chatbots such as ChatGPT and Gemini through Canva's MCP. TechCrunch only
  • Researchers from Leonardo.ai, which Canva acquired in 2024, contributed to the model. TechCrunch only
  • Canva is developing two cost-efficient versions, Craftfast and Craft-ultrafast, to be released soon. TechCrunch only
  • Canva cited better quality and speed control, in addition to cost, as benefits of using its own model rather than several third-party image models. TechCrunch only
  • Canva was on track for $4 billion in revenue last year but cut its 2026 growth forecast; its valuation reportedly fell from $42 billion to $35 billion according to external analysts. TechCrunch only

Context

Canva has launched Canva Craft, its first proprietary image generation model, positioning it against comparable models from Google, OpenAI, ByteDance and Meta. The company says the model plans a layout before generating an image, aiming for accurate object details, legible text and precise edits that leave other parts of the image unchanged.

The model is available to all Canva users and will come to the Canva API, with planned support in third-party chatbots through Canva's MCP. Two lower-cost variants are in development.

How the Agent Panel Scored This

Independent AI personas each scored importance 1–10; the story's score is a trimmed mean (highest and lowest votes dropped). Vote spread: 1.

4 News Editor Canva's launch of a proprietary image model is a concrete product release available to all users, but it is a company-level product announcement of limited broad societal impact, with the financial details being secondary.
4 Business Analyst Canva's launch of a proprietary image generation model is a meaningful product and strategic move in a competitive AI space, but it is a single company product release with limited market-wide impact rather than a market-moving event.
3 Informed Citizen A product launch from a single company adds a new in-house image model available to Canva users, a niche tech/product update with limited impact on ordinary people beyond existing Canva users.

Source

← All stories