Canva has launched Canva Craft, its first in-house AI image-generation model, developed by its research team to compete with tools from Google, OpenAI, ByteDance and Meta.

The model is designed specifically for creative workflows, with a focus on accurate layouts, readable text and precise editing rather than image generation alone.

Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram

A Two-Stage Approach To Image Generation

Craft first interprets a user’s prompt and creates a structured plan covering spatial relationships, scale, visual hierarchy and text placement. It then generates images at resolutions of up to 2K across 30 aspect ratios.

Canva says this approach improves object detail, text legibility and the ability to edit individual elements without disrupting the rest of a composition, helping creators maintain consistent layouts across different formats.

Trained On Design, Not Just Photography

Unlike models trained primarily on photographs, Craft was developed using human-designed templates from Canva’s library, including posters, social media posts, presentations, infographics, menus and typography-heavy layouts.

Images generated with Craft can also be converted into layered designs using Canva Magic, the company’s image-to-design tool. The feature builds on Canva’s earlier work on AI models that can interpret design elements and layers, supporting its ambition to offer an integrated creative platform.

Expanding Canva’s AI Ecosystem

Craft is available to all Canva users through the Dream Lab tab. The company also plans to offer the model through its API and enable image generation through third-party chatbots, including ChatGPT and Gemini, using Canva’s Model Context Protocol (MCP) integration.

Developing an in-house model could reduce Canva’s reliance on external image-generation providers, helping lower API costs while giving the company greater control over performance, speed and product quality.

Leonardo.ai Strengthens In-House Development

Canva’s acquisition of media-generation platform Leonardo.ai in 2024 has contributed to its AI expansion, with Leonardo researchers involved in developing Craft.

The company is also working on two lower-cost versions, Craftfast and Craft-ultrafast, which are expected to launch soon. These variants would offer faster, more economical options for everyday creative tasks alongside the main model.

Balancing AI Growth And Costs

The launch comes as Canva seeks to balance revenue growth with rising AI infrastructure costs. The company was on track to reach $4 billion in revenue last year but later lowered its growth forecast for 2026. According to external analysts, the revision contributed to a decline in its valuation from $42 billion to $35 billion.

Developing its own AI technology could help Canva manage costs while expanding its product offering. The move reflects a broader challenge across the software industry: delivering increasingly sophisticated AI features without allowing infrastructure expenses to undermine profitability.