A new AI image generator has surfaced, confidently declaring its superiority over existing models, including the well-known Midjourney. Initially shrouded in mystery under the codename “red_panda,” this innovative tool is officially named Recraft. It has recently topped the Artificial Analysis’ Text-to-Image Model Leaderboard, a measure of the leading text-to-image models based on human preference, igniting curiosity about its origins and functionality.
Earlier this week, the conversation around red_panda sparked intrigue, with many in the tech community anticipating details about its creators. PetaPixel had considered reporting on the emerging technology but deferred due to limited information. However, Recraft AI, a London-based platform led by Anna Veronika Dorogush, has now stepped into the spotlight, confirming its identity on social media. With a welcoming message that read, “Nice to meet you! We are red_panda, but friends call us Recraft,” the company revealed its groundbreaking creation.
Recraft AI has made significant waves in the industry, having raised $12 million earlier this year from venture capital investors, a clear indication of its promise. With the release of Recraft V3, the company boasts a revolutionary AI model that operates with a profound understanding of design language. According to their announcements, Recraft V3 offers unparalleled quality in text generation, surpassing established models from competitors like Midjourney and OpenAI.
More than just an image generator, Recraft V3 is positioned as a robust design tool capable of enhancing text placement, exerting style control, and achieving the highest quality standards currently available. As highlighted on their social media channel, this platform has set a new benchmark for excellence in image generation — a claim they back by offering a unique feature: the ability to generate long textual content on images.
Focusing on the needs of graphic designers, Recraft V3 emphasizes user control in text size and placement, allowing for the creation of detailed, professional-quality visuals ideal for branding, marketing, and intricate graphic layouts. Users can also explore various styles to ensure that image outputs meet their specific aesthetic requirements. For businesses looking to develop a cohesive brand image, Recraft offers the capability to refine style candidates until they perfectly embody the desired look and feel.
Among its arsenal of tools are capabilities such as background removal, a Generative Fill-like feature, outpainting, mockup generation, vectorization, and upscaling. Furthermore, Recraft V3 supports diverse styles, including photorealism, studio photography, HDR, hard flash, and black and white interpretations. During preliminary tests conducted by PetaPixel, Recraft showcased impressive skill, even ensuring accurate finger placement in images of photographers handling cameras.
Recraft is accessible to users with a free option that provides 50 credits daily, in addition to tiered paid packages: Basic at $10, Advanced at $27, and Pro at $48. Interested users can try Recraft by visiting their website or through the newly launched app, available on both iOS and Android platforms. As Recraft positions itself as a frontrunner in the AI image generation arena, the tech community awaits to see how it influences design practices and the future of branding and marketing imagery.