Controversy continues to surround the world of AI-generated imagery, even as AI-generated imagery is used in elections became a source of concern this week at the World Economic Forum, startups continue to plow the new furrow of AI tools for creators.
The latest is Recraft, an artificial intelligence graphic design generator aimed at professionals, which raised a $12 million Series A round led by Khosla Ventures in Silicon Valley, along with former GitHub CEO Nat Friedman. RTP Global, Abstract VC, Basis Set Ventures, Elad Gil and several other angel investors also participated.
Admittedly there are now countless Generative AI design tools such as Jasper, Adobe Sensei, Let’s Enhance and many more. However, Recraft claims it is among the first to be a “foundational” tool, in the sense that it builds its own Foundation Model – a pre-trained deep learning algorithm – to create consistent design elements such as icons and images. that can be modified and used within a brand’s style controls. It also claims to have amassed over 300,000 since its launch eight months ago.
But this isn’t designed as a tool to spit out funny images of cowboy dogs riding horses or the like. In addition to being able to produce these “raster images”, it can also create vector images that are infinitely scalable and used in professional graphic design realms, as opposed to platforms where image quality can often be limited.
Perhaps this is to be expected, since the founder, Anna Veronika Dorogush is not exactly a non-technical founder. She created CatBoost, a high-performance open source library for gradient boosting on decision trees, for example. Formerly head of machine learning systems with search engine Yandex in Moscow, Dorogush and her 13-member team are currently based in London.
She told me in a call that the fundraising would boost her efforts to build her own foundational model: “You have to give users a lot of control over the results … over the style, so you can have consistent images and control over things like colors of the brand or the level of detail, as well as the repeatability of the resulting image.”
“But if the model initially cannot produce a pool player or a ballet dancer, then this means that just using fine tuning is not enough to deliver high-quality output, as of now. So we have to build our own model and we are working on it right now,” he added.
He says Recraft is very much aimed at professionals: “What’s unique about what we do is that we don’t just provide image creation. We also provide style control: the ability to create your own style and then create images in your own style… This is important if you want to build a brand and develop it, create marketing materials, create consistent advertising.”
I asked her what the chances were that they would be sued by an artist whose style was used on the Recraft platform.
“We declare in ours terms that if an artist uploads something to the system as a sterile reference—like the image for which they can create new images—we don’t use them to train our model,” says Dorogush.
“We don’t use them to train our model.”
He added that Recraft effectively adds generic styles to images uploaded by artists: “You provide the style to the model. So you have a style reference. And then you create images using that reference style. So the user’s action is to provide the style.”
Dorogush also says that, rather than displacing graphic designers, tools like Recraft are more likely to simply make graphic design more accessible in arenas where uninspired stock images might have been used: “A year ago, if you were writing a children’s book , you had stock images. Now, the same people can have much more beautiful illustrations and images in the same books.”
Of course, Recraft isn’t the only startup playing in this space, and last year I covered how Berlin-based Kittl had raised a €10.8 million ($11.6 million) Series A for its design platform to enable users to turn ideas into graphic products to create professional-quality designs, quickly.
This week, CEO Nicolas Heymann told me on a call that products like Recraft are part of an increasingly sophisticated set of tools that provide alternatives for designers: “I think if Recraft is able to produce design elements successfully — similar to how image generation models can be a good replacement nowadays for stock photo libraries — then they can definitely be needed [some market] share of these icon libraries.”
Dorogush says there’s room for more than one player right now, and that Kittl has a different use case than Recraft: “We have a unique technology for creating consistent images with style, and that’s our main focus. We focus on graphic designers and marketers who need consistency in style, while Kittl focuses on merchandise and I have huge respect for them. It’s a great company.”
Whatever the case, it seems that investors are excited by the possibility of avoiding the controversy that has attracted some AI visualization and design platforms in favor of supporting these business-focused AI tools.
In a statement, Nikita Shamgunov, Partner at Khosla Ventures commented: “We are seeing rapid and significant transformation of the design space driven by genetic artificial intelligence. But to date, many of the production AI design solutions target consumers rather than professionals who require high degrees of control. Recraft offers professional workflows such as vector images, style controls and end-to-end content production, all powered by internal foundational models.”
