F13
AI startup developing foundation models for precise, editable vector graphics from text, images, or existing graphics.
Website: https://www.f13.com/
Cover Block
Open sources
| Field | Detail |
|---|---|
| Name | F13 |
| Tagline | AI startup developing foundation models for precise, editable vector graphics from text, images, or existing graphics [f13.com] [Tech.eu, September 2026] |
| Headquarters | Berlin, Germany [f13.com] [Tech.eu, September 2026] |
| Founded | 2026 [cfotech.co.uk, October 2026] |
| Stage | Pre-Seed [Tech.eu, September 2026] |
| Business Model | API / Developer Platform [EU-Startups / EUVC, September 2026] |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2): Gregory Janik and Ingimar Tomasson [cfotech.co.uk, October 2026] [The Next Web, September 2026] |
| Funding Label | Pre-seed |
| Total Disclosed | ~$5,000,000 [Tech.eu, September 2026] |
Links
Open sources
- Website: https://www.f13.com/
What an Investor Needs First
PUBLIC F13 is a Berlin-based AI startup building foundation models that generate precise, editable vector graphics from text prompts, images, or existing graphics, and it merits attention now because it came out of stealth in September 2026 with a $5 million pre-seed round behind a product thesis that is narrower, and potentially more commercially useful, than general image generation [Tech.eu, September 2026] [EU-Startups / EUVC, September 2026] [f13.com]. The founding story is relatively straightforward in the public record: Gregory Janik and Ingimar Tomasson previously worked together at Knownunity, then founded F13 to address the persistent gap between visually plausible AI images and graphics that remain structured, editable, and suitable for professional use [The Next Web, September 2026] [cfotech.co.uk, October 2026].
The product claim is that F13’s models produce SVG-style outputs that preserve layout control, proportions, editability, and data fidelity for workflows such as charts, scientific diagrams, maps, and brand-consistent documents, with early access available through a developer API and web platform [Vestbee, September 2026] [EU-Startups / EUVC, September 2026] [The SaaS News, September 2026]. That positioning matters because the differentiation appears to rest less on broad image quality and more on vector-native utility, which is a clearer wedge if the company can prove reliability inside design, documentation, and embedded software workflows [Tech.eu, September 2026] [f13.com].
The team looks relevant to the problem, though still early. Public reporting identifies Janik as co-founder and CEO, Tomasson as co-founder and machine-learning researcher trained at University College London, and multiple reports describe a wider bench spanning machine learning researchers with prior experience at Meta and Porsche, alongside designers who have worked with brands including Versace, Gucci, and Nike [The Next Web, September 2026] [ETIH EdTech News] [cfotech.co.uk, October 2026].
On financing, F13 has disclosed a single $5 million pre-seed round led by Credo Ventures and Point Nine Capital, with angel participation from Carles Reina and Jack Richardson, which is consistent with a developer-platform business model that will likely need time to convert technical novelty into repeatable API or software revenue [Tech.eu, September 2026]. Over the next 12 to 18 months, the public watchpoints are practical rather than conceptual: evidence of production adoption, proof that the API and web product can serve professional workflows at sufficient accuracy and speed, and any signal that F13 can move from launch coverage to named users or measurable commercial traction [EU-Startups / EUVC, September 2026] [Vestbee, September 2026].
Partially corroborated -- Core company, founder, and funding facts are corroborated by multiple public sources, but product performance and team-depth claims rely partly on company or startup-media reporting.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Pre-Seed |
| Business Model | API / Developer Platform |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (2) |
| Funding | Pre-seed, total disclosed approximately $5,000,000 |
Inside the Company
PUBLIC
What is established in the public record is narrow but usable: F13 presents itself as a Berlin-based lab training models that generate vector graphics natively, and its public site frames the product around accurate, editable SVG output rather than conventional pixel images [f13.com]. The available company materials do not surface a legal entity name, but they do support Berlin as the operating base and foundation-model research for vector graphics as the core mission [f13.com].
The company appears to have been founded in 2026, with Gregory Janik and Ingimar Tomasson identified as co-founders in later public reporting; the site itself, however, is the cleaner source for the company’s own positioning than for incorporation detail [f13.com]. On chronology, the first clear milestone visible from public company materials is the launch of its web presence with a waitlist and product demonstrations centered on production-grade vector graphics generated from prompts [f13.com].
A second milestone is the company’s public emergence in September 2026, when multiple outlets reported a $5 million pre-seed round and described F13 as coming out of stealth, but those financing details sit outside the source set permitted for this section and are therefore better treated elsewhere in the report. For the purpose of company overview alone, the conservative read is that F13 moved from a minimally descriptive web presence to a named Berlin startup with a defined vector-graphics thesis during 2026, while key corporate details remain limited in company-controlled public materials [f13.com].
Partially corroborated -- This section relies primarily on the company website, which supports headquarters and product positioning, but does not independently confirm legal entity detail or full founding chronology.
Under the Hood
Reported and inferred
F13 is making a fairly specific technical bet: that image generation for professional graphics is less about photorealism and more about structure, editability, and precision. Public materials describe the product as a vector-native model stack that generates editable graphics from text prompts, images, or existing graphics, with outputs intended to remain usable inside professional design and documentation workflows [f13.com] [Tech.eu, September 2026] [EU-Startups / EUVC, September 2026]. The company website says its models "generate production-grade vector graphics" and return "an accurate, editable SVG in seconds," while press coverage frames the same core idea as foundation models for precise vector graphics rather than pixel-based image synthesis [f13.com] [The Next Web, September 2026].
The practical distinction matters if the model can preserve object relationships and editable structure rather than flattening everything into an image. Coverage in Vestbee and EU-Startups / EUVC says F13 is targeting use cases such as data visualizations, scientific diagrams, maps, and brand-consistent documents, and that its system is meant to preserve proportions, layouts, editability, and data accuracy in SVG-style outputs [Vestbee, September 2026] [EU-Startups / EUVC, September 2026]. That positioning suggests a developer and design tooling wedge, reinforced by reports that F13 is offering early access through both a developer API and a web platform, although no verified public demo, benchmark, or customer deployment was identified in the available sources [EU-Startups / EUVC, September 2026] [Tech.eu, September 2026].
Partially corroborated -- Product scope is corroborated across the company website and multiple September 2026 publications, but performance claims and workflow usability remain largely company-described and not independently benchmarked.
Market Research
PUBLIC The market matters now because generative AI has made image creation common, but professional design and document workflows still need outputs that stay editable, structurally accurate, and usable after generation, a gap F13 is explicitly targeting [Tech.eu, September 2026] [EU-Startups / EUVC, September 2026] [Vestbee, September 2026].
Public reporting does not provide a direct TAM, SAM, or SOM for vector-native generative graphics, so the more defensible way to frame the opportunity is through adjacent software and content-creation budgets rather than a precise top-down market claim. The cited coverage places F13 at the intersection of developer tools, design software, and enterprise document generation, with use cases spanning data visualizations, scientific diagrams, maps, and brand-consistent documents [EU-Startups / EUVC, September 2026] [The SaaS News, September 2026]. That is a meaningful clue about budget ownership: if adoption occurs, spending could come from teams that already pay for design suites, data visualization tooling, mapping workflows, or document automation, rather than from a stand-alone "AI image" budget.
The demand driver is not image novelty, it is workflow usability. Multiple reports make the same point in different language: existing image generators are strong at pixels but weak at producing SVG-style assets that preserve proportions, layouts, editability, and data accuracy [Vestbee, September 2026] [The SaaS News, September 2026]. That matters most in categories where a user needs to revise, localize, resize, or brand an asset after generation, which helps explain why the reported early product surface is a developer API and web platform rather than a consumer image app [EU-Startups / EUVC, September 2026].
The substitute set is broad, even if named competitors are not confirmed in the source base. In practice, F13 is competing against manual vector design workflows, existing design software with AI features, general-purpose image models that export raster assets, and internal tooling built by teams with recurring needs for diagrams or charts [Tech.eu, September 2026] [Vestbee, September 2026]. The adjacent markets are therefore larger than "AI vector graphics" as a narrow label would suggest, but they are also more crowded, because buyers can often solve the problem imperfectly with incumbent design products or human workflows.
Macro and regulatory forces are present, though the source set supports only a cautious read. On the macro side, the broader enterprise push toward content automation supports experimentation with generated diagrams, maps, and branded visual assets, particularly if they can be integrated through an API into existing software stacks [EU-Startups / EUVC, September 2026]. On the constraint side, F13's own positioning around accuracy and factual utility implies that trust will matter more here than in consumer image generation, especially for scientific, cartographic, or data-linked outputs where an attractive graphic can still fail a professional review if the structure is wrong [Vestbee, September 2026] [LinkedIn, retrieved 2026].
| Market lens | What the public evidence supports | Sources |
|---|---|---|
| Core category | An emerging vector-native generative graphics segment focused on editable SVG-style outputs | [Tech.eu, September 2026] [Vestbee, September 2026] |
| Budget adjacency | Design software, visualization tooling, mapping, and document-generation workflows | [EU-Startups / EUVC, September 2026] [The SaaS News, September 2026] |
| Primary buyer motion | Early access via developer API and web platform suggests both product-led experimentation and embedded workflow use cases | [EU-Startups / EUVC, September 2026] |
| Main demand driver | Need for precise, controllable, editable outputs rather than static AI-generated images | [EU-Startups / EUVC, September 2026] [Vestbee, September 2026] |
| Key adoption constraint | Professional users likely need factual and structural reliability before production deployment | [Vestbee, September 2026] [LinkedIn, retrieved 2026] |
The picture here is attractive but still early. The market signal is clearest at the problem level, namely that professional teams need editable machine-generated graphics, while the size of the spend pool remains unquantified in the available public reporting.
Partially corroborated -- Section relies on named public coverage from Tech.eu, EU-Startups / EUVC, Vestbee, The SaaS News, and LinkedIn, but no third-party market sizing report or confirmed numeric TAM data was identified.
Competition and Substitutes
MIXED F13 is positioning itself less against another vector design tool than against the current failure mode of general-purpose image models, which can create attractive pictures but often struggle to produce editable, geometry-faithful graphics for professional use cases [Tech.eu, September 2026] [Vestbee, September 2026].
The competitive map is therefore best read across three adjacent segments, even though the available source set does not name direct startup rivals. First are general image-generation platforms, the broad AI category F13 is implicitly reacting to when it argues that existing systems do not reliably output usable SVG-style assets for diagrams, maps, and data visuals [Vestbee, September 2026] [The SaaS News, September 2026]. Second are incumbent vector and design workflows, where professionals still rely on manual or semi-manual creation because editability, proportion control, and data accuracy matter more than visual novelty [EU-Startups / EUVC, September 2026] [f13.com]. Third are adjacent developer tools, since F13 is offering early access through an API and web platform rather than only through a designer-facing application, which suggests competition for technical workflow integration as much as for creative seat licenses [EU-Startups / EUVC, September 2026].
Where F13 appears strongest today is in focus. Public reporting is unusually consistent on the same point: the company is building foundation models specifically for vector graphics, with outputs meant to remain precise, controllable, and editable across use cases such as scientific diagrams, maps, and brand-consistent documents [Tech.eu, September 2026] [EU-Startups / EUVC, September 2026] [f13.com]. That specialization could matter if buyers value structural correctness over stylistic range, and if the product reaches developers through an API before larger horizontal AI vendors treat vector-native generation as a serious product surface [EU-Startups / EUVC, September 2026]. The edge is still perishable. A focused model thesis can be durable when it rests on proprietary training data, workflow entrenchment, or unusually strong evaluation loops, but none of those moats is yet publicly demonstrated in the current record [Tech.eu, September 2026] [The Next Web, September 2026].
The company is most exposed where horizontal platforms and incumbent design ecosystems already control user attention and distribution. Even without a named startup competitor in the source set, the practical benchmark is the broader class of image-generation products that already have large user bases, mature interfaces, and substantial compute budgets, even if their vector outputs remain limited today [Vestbee, September 2026] [The SaaS News, September 2026]. F13 also does not yet appear, on public evidence, to own a downstream channel such as a dominant design suite, a large open-source community, or named enterprise design accounts that would make replacement costly [EU-Startups / EUVC, September 2026] [Tech.eu, September 2026]. That leaves the company dependent on product quality and developer adoption at a stage when distribution can matter as much as model performance.
The most plausible 18-month scenario is a split market rather than a winner-take-all outcome. F13 is the likely winner if vector-native accuracy proves hard for general image models to replicate and if developer workflows around charts, diagrams, and brand assets start to prefer API-based generation over manual design steps [EU-Startups / EUVC, September 2026] [f13.com]. The likely loser in that case is the status quo of pixel-first AI for structured graphics, because the gap F13 is targeting is specifically the need for editable SVG-style output with preserved layout and proportions [Vestbee, September 2026]. The reverse scenario is also straightforward: if horizontal AI platforms make vector generation a bundled feature quickly enough, then a standalone specialist like F13 could find its wedge narrowed before it establishes clear distribution or customer lock-in [Tech.eu, September 2026] [The Next Web, September 2026].
Partially corroborated -- This section relies on multiple public sources for F13's positioning and product claims, but the source set names no direct competitors, so the competitive map is partly analytical rather than directly corroborated by company-by-company public reporting.
Opportunity
PUBLIC
The prize here is not a nicer design tool, it is the chance to become core infrastructure for any workflow that needs machine-generated visuals to remain editable, precise, and usable inside professional software stacks [Tech.eu, September 2026] [EU-Startups / EUVC, September 2026] [f13.com].
The headline opportunity is straightforward. If F13 can make vector generation reliable where general image models remain weak, it could become the default model layer for charts, diagrams, maps, and branded document graphics across software products and enterprise workflows [Vestbee, September 2026] [The SaaS News, September 2026]. That outcome is reachable, not merely aspirational, because the company is already positioning around a specific technical gap rather than a broad creative AI claim: multiple reports describe outputs as vector-native, editable, and suitable for professional use cases, while the product is already being offered through an early-access API and web platform [EU-Startups / EUVC, September 2026] [f13.com] [Tech.eu, September 2026]. In practical terms, infrastructure status matters more than consumer virality here. If developers can call F13 to generate SVG-style assets that preserve layout and proportions, adoption can travel through product teams, document systems, analytics platforms, and design workflows without F13 needing to own the full end-user application [Vestbee, September 2026] [EU-Startups / EUVC, September 2026].
The plausible paths to scale are narrower than they are in horizontal generative AI, but each path has real operating use if the product works as described.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Embedded visual infrastructure | F13 becomes the API layer that software teams use when they need generated charts, diagrams, maps, or branded vector assets inside their own products | A public launch of the developer API and repeatable integrations into SaaS workflows | The company is already offering early access through a developer API and web platform, and reported use cases map cleanly to software-embedded generation rather than only standalone design work [EU-Startups / EUVC, September 2026] [Tech.eu, September 2026] |
| System of record for professional AI graphics | F13 wins design, knowledge-work, and document workflows where editability and brand control matter more than photorealism | A model release that consistently preserves proportions, layouts, and editability at production quality | Coverage repeatedly frames the wedge as precise, controllable, editable vector output for professional workflows, including data visualizations and brand-consistent documents [Vestbee, September 2026] [EU-Startups / EUVC, September 2026] [f13.com] |
| Category leader in technical visuals | F13 becomes the specialist model provider for scientific diagrams, maps, and structured business graphics, a segment general image models have not served well | A category-tipping customer or platform partner in analytics, education, or enterprise productivity | The cited use cases are unusually specific for an early company, and they point to domains where factual structure and editable output matter more than open-ended image generation [EU-Startups / EUVC, September 2026] [The SaaS News, September 2026] |
The compounding logic rests on workflow fit. A vector model that performs well in one production context should generate better training signals from adjacent professional tasks, because charts, maps, diagrams, and brand templates all reward structural correctness, not just aesthetic plausibility [Vestbee, September 2026] [f13.com]. If the API is adopted first by developers, each implementation can create recurring prompt patterns, failure cases, and domain-specific graphics data that help improve output quality over time, which in turn makes the product more useful for higher-stakes workloads. The evidence for that flywheel is early, but it is visible in how the company describes the product surface: native vector generation, production-grade SVG output, and an API plus web workflow, rather than a single consumer interface [f13.com] [EU-Startups / EUVC, September 2026]. A second layer of compounding could come from editability itself. Once generated graphics enter existing design and document systems as usable vector assets, switching costs rise because teams can incorporate the output directly into branded and iterative workflows instead of treating it as disposable imagery [Vestbee, September 2026] [The SaaS News, September 2026].
The size of the win is difficult to pin down from public evidence because no market sizing source or direct public comparable is confirmed in the source set. Even so, the strategic value can be framed with some discipline. If F13 becomes the specialist infrastructure provider for AI-generated vector graphics inside software and enterprise content workflows, the company could plausibly support an outcome measured in the high hundreds of millions to low single-digit billions of dollars in enterprise value, especially if the API becomes embedded across multiple categories of productivity and design software (scenario, not a forecast). That is an analytical scenario rather than a sourced valuation claim, and it rests on one central assumption supported by the reporting: that precise, editable graphics remain a distinct technical gap versus pixel-based image generation, and that F13's first-model launch converts that gap into durable developer adoption [Tech.eu, September 2026] [Vestbee, September 2026] [EU-Startups / EUVC, September 2026].
Partially corroborated -- This section relies primarily on startup publication coverage and the company website; core product and funding claims are corroborated across multiple public sources, but scale outcomes and valuation framing remain analytical scenarios rather than independently verified market facts.
Sources
Open sources
[f13.com] F13 | https://www.f13.com/
[Tech.eu, September 2026] F13 emerges from stealth with $5M for vector graphics AI | https://tech.eu/2026/09/23/f13-emerges-from-stealth-with-5m-for-vector-graphics-ai/
[cfotech.co.uk, October 2026] Gregory Janik and Ingimar Tomasson founded F13 | https://cfotech.co.uk/
[The Next Web, September 2026] F13 raises $5M pre-seed to build an AI model for precise vector graphics | https://www.thenextweb.com/news/f13-raises-5m-pre-seed-vector-graphics-ai
[EU-Startups / EUVC, September 2026] F13 raises $5m to build precise vector-graphics AI | https://www.eu.vc/p/f13-raises-5m-to-build-precise-vector-graphics-ai
[Vestbee, September 2026] F13 launches from stealth with $5M to tackle AI’s visual accuracy gap | https://www.vestbee.com/insights/articles/f13-lunches-from-stealth-with-5-m
[The SaaS News, September 2026] F13 Raises $5M Pre-seed | https://www.thesaasnews.com/news/f13-raises-5m-pre-seed/
[ETIH EdTech News, retrieved 2026] F13 team background coverage | https://www.edtechinnovationhub.com/
[LinkedIn, retrieved 2026] Tomas Fanta, CFA - Venture investor, CEE&Baltics, AI, Blockchain | LinkedIn | https://www.linkedin.com/in/tomas-fanta/
Articles about F13
- F13's AI Model Aims for the Precise, Editable SVG — The Berlin startup, backed by Credo and Point Nine, is building foundation models to generate editable graphics for data, science, and design workflows.