Recraft builds image-generation AI models from scratch and competes with Midjourney, DALL‑E, Adobe Firefly, Stable Diffusion and Black Forest Labs. Its models are especially aimed at generating brand-compliant images, enabling precise logo placement and creating marketing materials that follow brand guidelines. Recraft’s V3 model, codenamed “red_panda,” topped the Artificial Analysis benchmark last year, beating DALL‑E and Midjourney. The company says it recently passed $5 million in ARR and 4 million users. That positioning brings Recraft closer to competing with design tools like Canva for branding-focused content generation. The company is led by founder and CEO Anna Veronika Dorogush, who previously worked on machine learning at Yandex and held roles at Google and Microsoft. Recraft builds a foundational AI model to generate professional-grade graphic design assets with style control, including vector and raster images that can be scaled and tweaked for brand consistency. The product targets graphic designers and marketers who need consistent style, brand-color control, and the ability to iterate on generated assets. Founder Anna Veronika Dorogush, creator of the CatBoost open-source library and former head of machine learning systems at Yandex, leads a 13-strong team based in London. Recraft says it has amassed over 300,000 since its launch eight months ago. The company emphasizes generating scalable vector images for professional workflows rather than consumer-oriented novelty images. Recraft plans to use its recent funding to build its own in-house foundation model to improve coverage and control over generated subjects and styles.

Total raised

$42M

Funding rounds

2

Latest round

Series B

Latest activity

May 2025