.webp)
F13 has emerged from stealth with a USD $5 million pre-seed round led by Credo Ventures and Point Nine.
The London-based AI company is also unveiling a model that generates vector graphics, a format widely used in design and engineering because files remain editable and can scale without losing quality.
Its system is designed to produce visuals that are both technically accurate and visually polished. F13 argues that general-purpose image models can create striking pictures but often struggle with tasks that require exact proportions, accurate data representation and adherence to design rules.
That gap is central to the company's pitch. According to F13, current AI tools often fail to produce charts that accurately reflect underlying data, scientific diagrams with the right proportions, or branded materials that follow a fixed design system.
Users can prompt the model with text, images or existing graphics. It then generates editable vector outputs, allowing designers and other teams to refine the work rather than start again.
F13 has launched to early-access customers and says current use cases include data visualisations and charts, editorial infographics, scientific and educational diagrams, maps, and brand-consistent presentation and marketing layouts. It is offering the model through an API and a web-based platform.
The startup was founded by Gregory Janik and Ingimar Tomasson, who met while working at Knownunity on efforts to integrate precise visuals into its product. Janik is a former founder in machine learning infrastructure and financial technology, while Tomasson is a machine learning researcher with a background in mathematics.
The wider team includes machine learning researchers and designers with previous experience at companies including Meta and Porsche, and brands including Versace, Gucci and Nike.
Technical focus
F13 says its model is built around technology intended to improve how AI understands and works with two-dimensional space. It says this gives the system more control and precision in producing vector graphics than broader image and code-generation models.
In internal testing against general-purpose models, F13 says its system outperformed offerings from larger AI labs in generating vector graphics with accuracy and clarity. The company did not provide benchmark data in the announcement.
The financing also included angel investors such as Carles Reina of Baobab Ventures and Jack Richardson of mainframe. F13 says the new capital will be used to bring its first model to market and expand compute, talent and data resources.
Investors framed the bet as one on a specialist approach rather than a broad image-model strategy. They pointed to the volume of visual material companies need to produce and the limits of existing AI tools in work that depends on exact layouts, proportions and brand consistency.
"We didn't want to build another model that's impressively expressive and quietly wrong. We create visuals that are not only beautiful to look at, but are true to the brief, with the correct proportions, accurate data, and precise layouts. That's the only version of this technology that's genuinely useful to discerning professionals who know what good looks like and are doing high-stakes design work. General-purpose AI models aren't coming close to us yet, because they weren't built for this job and have fundamental architectural limitations. This space needs a level of specialisation the big labs can't justify right now. Our first cheque came from a chance conversation with an investor, back when F13 was nothing more than an idea. We raised the rest of the round without ever putting together a pitch deck. This is a testament to the size of the problem and the strength of our team, and we're proud to have such respected investors behind us as we become the first to solve it," said Gregory Janik, co-founder and chief executive officer of F13.
Matěj Míček, tech partner at Credo Ventures, outlined the firm's reasoning. "Gregory and Ingimar have a strong track record of finding solutions to deeply technical problems, but what sets them apart is taste. Together, they bring state-of-the-art machine learning research, deep roots in the fashion and design world, and a proven history of founding and scaling companies. We believe frontier models that understand 2D space precisely and have a real sense of aesthetics will be a significant part of the next wave of the AI revolution. We've made a bet that Gregory and Ingimar are the right people to win in this space," he said.
Point Nine made a similar argument about demand for more accurate visual generation in business settings. "Every company deals with a tremendous amount of visuals: graphic design, charts & diagrams, marketing assets, etc. And yet this is exactly what today's AI still gets wrong. There is a reason why software engineers are using agents for everything, but designers still rely on manual workflows. At Point Nine we believe F13 will change this," said Pawel Chudzinski, partner at Point Nine.