On a train from Ostrava to Prague, Jiří Třečák and his teammates decided their pitch deck needed something slides could not provide: proof. Before they reached their meeting with Credo Ventures, they assembled a six-screen app made for the investor. Třečák later remembered the sprint as taking about ten minutes. The detail is small, almost comic, and it contains the founding instinct behind Supernova. If a picture could become a working interface faster, the tedious distance between imagining software and building it might collapse.
That instinct came from repetition. Before Supernova, Třečák had helped create more than 70 mobile apps and websites as an engineer, product builder, employee, and freelancer. Every project passed through the same translation loop. Designers made decisions visually. Developers reconstructed them in code. Product managers carried requirements between the two. Details fell out along the way, came back as questions, and consumed hours that nobody had entered the profession to spend.
Třečák was born in Ostrava and had been writing software since childhood. He studied applied informatics and information technology, then worked across mobile engineering and product roles, including STRV and the Amsterdam startup StartMonday. A former colleague remembered him packing a unicycle and moving to Amsterdam while that company was still only an idea. The image fits: technically curious, comfortable with a leap, and willing to bring one odd piece of luggage.
A demo that met the real world
Supernova launched in 2018 around a direct promise: convert design into production code. It was not a toy idea. The team spent three years on research and development, building technology that could interpret a visual composition rather than merely export a crude template. Flexible solo creators could use it. Thousands tried it. The company reached Y Combinator's Winter 2019 batch after an earlier application had failed, going from Ostrava to Silicon Valley without first taking the conventional detour through Prague.
Then the clean promise encountered established products. A freelancer beginning with an empty file can adjust a workflow around a new tool. A large company cannot casually discard its component library, naming conventions, accessibility rules, brand history, release process, or old code. Generated code might resemble the intended interface and still be unusable by the people responsible for maintaining it.
“We learned something critical: speeding up one step wasn't enough to deliver true value.”Jiri Trecak, reflecting on the Y Combinator period
That sentence describes a useful kind of founder disappointment. The mechanism worked, yet the workflow did not. The company could have polished the visible trick and kept selling speed. Instead, Třečák and his co-founders treated the missing information as the product. Teams needed design, code, documentation, tokens, components, and product decisions in one connected place. Supernova started over as a design-system platform.
The infrastructure years
Design systems sound tidy. In practice, they are living agreements between people who use different tools and are rewarded for different outcomes. A color token changes in Figma. A component in code still expects the old value. Documentation describes yesterday's behavior. Someone opens a ticket. Someone else asks which source is authoritative. The system designed to create consistency becomes another place where coordination can fail.
Supernova made that connective tissue its focus. The platform synchronized design tokens, component libraries, documentation, and code delivery. A $4.8 million seed round in 2022 supported a next-generation token manager and deeper enterprise work. Customers publicly named by the company grew to include Mozilla, Paramount, and Skyscanner; later reporting also identified KLM, Air France, Mews, and Mayo Clinic.
The shift demanded patience while faster-looking AI products arrived with spectacular demos. Třečák acknowledged the ego test of watching other startups advance more quickly. At one difficult point, Supernova had only weeks of runway left. But the supposedly slower work was accumulating something generic generators did not possess: structured knowledge of how a particular organization builds.
A shortcut becomes infrastructure when it remembers the constraints that make the work belong.
The founder stayed unusually close to that work. Třečák has described himself as an introvert who must perform extroversion through days of meetings. Coding is how he recharges. He once imagined reserving half a day each week to write it, partly as technical practice and partly as therapy. It is a revealing tension for an engineering CEO: the job pulls him away from the activity that first showed him the problem.
“Inside, I am still an introvert who recharges by sitting at a computer and writing piles of code.”Jiri Trecak on staying close to the craft
His public persona mixes systems thinking with open enthusiasm. On LinkedIn he calls himself a well-organized, want-to-know-everything geek. He has spoken at Google Flutter Interact, developer conferences, universities, and design-system events. Former colleagues describe someone who sees the full pipeline, from concept and user experience to engineering and marketing. That range helps explain why Supernova expanded rather than narrowed when its first answer proved incomplete.
Returning to the original promise
By 2024, Supernova was previewing Forge, an attempt to generate frontend work with a company's own design system in view. The principle was different from a blank-slate coding assistant. Precision came from limits: the approved tokens, the real components, the codebase, the commit history, and the guidance that teams had already agreed to follow.
In September 2025, the company brought that logic into Supernova 3.0 and Supernova Portal. The new workflow begins with an idea, supports parallel exploration and feedback, generates product requirements from the evolving prototype, and can pass specifications or code toward development tools. Product context travels through the process rather than being pasted into every prompt. For enterprise clients, Supernova offered help adapting existing design systems and code libraries so outputs matched the product already in use.
The funding gave the company resources to hire, expand awareness, and pursue customers in the United States and Asia. Supernova's team remained deliberately split. Třečák moved to the United States with his wife in 2021, and the company placed sales and marketing there while keeping much of development in Europe. Cofounder Oskar Kořistka led business and operations, Artem Ufimtcev led technology with the development team, and Václav Luštěk remained part of the founding group.
The geography mirrors the product. One side lives close to the market; the other stays close to the machinery. Třečák has been frank about the economics: European engineering is less expensive, and a growing Czech company can compete for talent differently than it could against the largest American technology employers. If the ambition is global, he has argued, the company must also sell where many global buyers are.
The people around the company also trace its route. Credo Ventures backed the early team. Y Combinator supplied an American network and a hard product lesson. EQT Ventures, Wing Venture Capital, Kaya VC, Reflex Capital, and J&T Ventures joined across later rounds, while Taiwania Capital led the Series A. Třečák's named co-founders cover product, operations, and technology. The pattern is less lone inventor than assembled system: people with distinct functions connected around one stubborn problem, much like the platform they built.
What seven years bought
Supernova's long route offers a corrective to the startup story built around instant revelation. The train demo mattered because it made an idea concrete. The years after it mattered because they exposed what the demo concealed. Třečák did not discard the original frustration. He kept changing the company's understanding of where that frustration lived.
At first, it seemed to live in the act of coding a design. Then it appeared in the handoff between disciplines. Finally, it widened into organizational memory: the accumulated decisions that let a team recognize whether new work is truly theirs. AI made generation faster, but it also made that memory more valuable. When anyone can produce a plausible interface, fit becomes the harder problem.
Třečák's aspiration is still recognizably the one that began Supernova: remove the mundane translation so people can spend more time on creative judgment. The language has expanded from designers and developers to product managers and stakeholders, from one conversion to the arc between idea and delivery. The company now calls its destination an artificial product team. The useful phrase inside that ambition is not artificial. It is product team. The work has to preserve how people decide together.
There is playfulness in the way Třečák describes finally testing the integrated system: an idea refined in real time, turned into a prototype, specifications, tickets, and even the outline of a business case. His reaction was less a polished executive statement than a surprised builder's exclamation. After years spent constructing the context layer, the old magic trick worked differently. It no longer skipped the mess around professional software. It had begun to carry the mess with it.