A player prompts for the game asset they want and gets it back rigged, balanced and ready to play. We already have it running inside a finished RTS.
M1A2 · USA
Su-57 · Russia
AH-64 · USA
Leopard 2 · EUGameGen is the creation layer: a player types a prompt and gets back a unit that is rigged, balanced and on-model, playable in seconds. DoctrineStrike is the first title running on it.
The game is finished. Over 190 units, every one of them generated by this pipeline, inside the real build. Everything below is captured footage, not a mockup.
About €2M for an 18-month runway to launch and the numbers that unlock a Series A. Roughly half of it non-dilutive, on top of the 4% ZEC tax regime.
Steam Next Fest in October, launch right behind it. Roblox proved players will build and AI made building programmable. Nobody holds the middle yet.
One of gaming's most storied RTS franchises built one of its most loyal communities, and it is still waiting for a true successor. Twenty-three years on, the fans keep the classics alive with mods.
Zero Hour is the third most-visited game on ModDB out of 78,355, twenty-three years after release. Its biggest mod has 2.4 million downloads and still took 39 of them on the day this was measured. Nobody is being paid for any of it.
Not one franchise. A whole genre: sixteen classics, and the last new modern military base-builder shipped in 2010. The demand never went anywhere.
Every one of those numbers is somebody doing unpaid expert work to keep a 2003 game alive. That is demand paid for in the most expensive currency there is: time. And it is locked behind professional tools, which is the exact lock GameGen removes.
Unconstrained output drifts off style, off scale and off the rules. That is why no shipped title lets its players generate.
The gap: a community waiting for its game, players wanting to create, and no title that delivers either, let alone both.
Concurrent players and franchise dates verified Jul 2026 (Steam, SteamCharts, CnCNet). Last new title excludes remasters and re-releases.
DoctrineStrike began as the game we were waiting for ourselves: a finished, classic-style RTS. Then we went further: an in-game workshop where any player prompts units, skins and whole factions, and GameGen returns them rigged, balanced and on-brand. What a modder did in weeks, a prompt does.
Same model, orders of magnitude more users, the moment generation becomes a game mechanic.
Prompt rails, an AI referee and balance-locked stats mean player content always fits the game. The safety layer is the product; the moat section shows it running live.
Five bounded templates (infantry, tanks, aircraft, buildings, weapons) and a top-down RTS camera keep player creations readable and safe to ship. It is the friendliest possible arena for consumer-scale generation.
ComfyUI is a brilliant engine inside a clumsy product: a separate app, inconsistent output, experts only. We closed that gap. Every stage is a clean drop-in slot inside Unity, driven by one shared style. Prompt → sprite → mesh → rigged prefab, scaled to its class and wired to play.
body
rotor · spins
gun · auto-aim
missiles · fireThe market already exists: Roblox and Fortnite pay player-creators billions a year. But those players only assemble pre-made blocks, because nobody lets them generate. We do, and every generation is a paid transaction the studio shares in.
Unreal made Epic a company. Fortnite skins made it a giant. A free game turned cosmetics into one of the biggest money machines in entertainment:
Fortnite revenue since launch, almost all of it cosmetics.industry estimates · directional
Price of the game. Free-to-play, funded entirely by skins & the battle pass.
Skins in the catalog, every one hand-made by Epic's art team.
It rejected a reboot that replaced the finished game with pay-to-win. DoctrineStrike ships the complete RTS for free and sells only player-generated cosmetics, with combat stats balance-locked: paying never buys power. That is the Fortnite lesson the cancelled reboot missed.
The genre's defining classic stayed alive 25 years on mods, without ever getting a remake. We rebuilt that style of RTS and replaced “wait for a modder” with “prompt it yourself.”
The face-scan in Where Winds Meet turns a photo into your in-game character. One clip passed 21M views since March 2026 — from a single cosmetic feature. DoctrineStrike regenerates the whole unit: weapon, skin, faction.
Fortnite runs on V-Bucks, and V-Bucks built Epic. DoctrineStrike mints no coin of its own: every price is denominated in partner credits, so every player wallet is an account with the partner. The studio earns a share of every generation.
Players sign in with the partner inside the workshop and buy credits; the studio earns a revenue share on every credit spent in-game. Quality tiers (low → ultra) and one-tap revert both trigger fresh generations: the more players create, the more everyone earns.
Players share, buy and sell their creations, the Roblox/UEFN model applied to generated content. Every resale is another paid generation, and the creator premium splits between the creator and the studio.
GameGen can license into other titles. Each new game brings new players and a new revenue line, and every new player brings more generations. The loop compounds.
Raw generation cost of a complete building: concept art, damage states, rubble, wired into the game.measured on the live pipeline · July 2026
A full seven-building faction, destruction included. The deepest hook in the workshop costs less to serve than a AAA skin costs to make.
Of raw compute buys roughly 7 to 14 complete factions. Price them like the games they replace and the margin is software, not gaming.
GenAI-for-3D by 2029, up from $1.9B in 2024 (~31% CAGR).The Business Research Company · directional
UGC-gaming market by 2032, the category GameGen sits inside.6Wresearch · directional
An image generation costs pennies; a full rigged 3D unit costs more, scaling with detail. Per player it stays tiny, and the volume is massive.
DoctrineStrike ships an in-game Level & Faction Editor where players generate their own tanks, units and whole factions, behind a partner sign-in. Every creator-player becomes a customer of the partner: modding turned into one click.
Creations spread through a community hub: a Reddit-style feed of prompts and result shots. Sharing costs nothing. Fielding a shared unit runs its prompt again, and that run is a fresh paid generation.
The deepest hook: building a whole faction, your own country's army included. Every unit, building and weapon is generated to match, with stats cloned from an existing faction so balance never breaks. A committed creator can spend more on one faction than on a full-priced game.
Generated for this pitch ·four player factions from one base tank, one prompt each
Oplot · Ukraine
Osório · Brazil
Jaguar · Mexico
Harimau · IndonesiaOne more run each and the faction has a face · emblem and key art, same pipeline, same style as the game's official factions




Everyone approaches this from one side. The middle seat — a shipped game where players generate safely and every generation is billed — is empty.
Meshy, Tripo, Luma: strong generation, sold as a tool. No game, no players, no distribution. They can copy the model; they cannot copy a finished title.
Roblox and UEFN own player creation at scale, and it is all pre-made blocks. Letting raw generation in would break their games; constraining it is not what their stack was built for.
World-model games dream frames on a server farm. DoctrineStrike is a real engine build that runs on today's hardware, where AI generates the content, not the frames. Different genre, different market, no head-on rival.
First with all three: a finished title, a safety layer and a credit economy. That combination is the category.
The first question every platform asks about player-generated content: “how do you stop the bad stuff?” Two layers, neither of them a filter bolted on afterwards.
Tripo auto-rigs the mesh; a hidden prompt forces three-quarter view, Z-axis and a fixed scale per class, so the result snaps into the slot already combat-wired.
Transform-style tweaks (position, rotation, scale), quality tiers low → ultra and a one-tap revert, all bounded so a player can refine the build but never break it.
Before a creation enters the game, an AI check answers two questions: is it the class this slot expects, and is it safe? Prompt a cat in the tank slot and it never reaches a match.
Custom units clone their combat stats from an existing unit; the player restyles the look, never the numbers. Nobody prompts a tank that shoots across the whole map, so creativity stays unlimited and the match stays fair.
Teams pitching this thesis usually bring mock-ups. We bring DoctrineStrike, a finished RTS whose entire art library is AI-generated. Every unit below was generated from a prompt, in one coherent style, then rigged and wired to play:
M1A2 · USA
T-14 · Russia
Leopard 2 · EU
Arjun · India
F-22 · USA
Su-57 · Russia
J-20 · China
Typhoon · EU
AH-64 · USA
Karrar · Iran
BMPT · Russia
Rafale · EUPlayable units and buildings generated, plus every icon, projectile and prop. 6 factions, one style.
Prompt → playable unit, fully automated. Built for any model as the default backend: drop-in, no rebuild.
Working demo and full walkthrough: this site, plus the in-engine editor.
The pipeline is already built around clean slots, so the partner needs no custom integration. Four stages (Text→Image, Image→Image, Image→Mesh, Text→Mesh) are wired and waiting. Switching the backend is a single selection. No code, no rebuild.
The game ships either way. With a sponsor: free-to-play. Without: premium launch. The partnership is upside, not a dependency.
The partner's model becomes the generation engine of a launch-ready RTS. Integration is one asset in the Workflows library, not a project.
Like “Made with Unity”: a credit on a real published title (store page, trailers, credits). Proof that your tech ships finished games.
Style injection keeps every image the partner returns on-model, which means fewer retries and a lower cost per asset.
The product manufactures its own marketing: every generation is a before-and-after moment built to be filmed. The face-scan clips in the market section carried one feature past 21M views. Our whole game is that feature, times every unit, skin and faction a player can imagine.
Steam Next Fest in October 2026 is the first public demo. Wishlist campaigns start now, and the demo is the conversion engine.
Seed credits put the first thousand creators to work before launch. Their builds fill the prompt exchange, and every viral clip converts into paid generations.
Launch rides both channels. Every player build doubles as an ad for the game and for the partner, so both sides fund the push.
Two people build the game. The channels run on an in-house AI marketing stack, agent-operated with human sign-off.
Who is behind this: two people, in months. Everything on this page, the 190+ units, the pipeline and the live demos included.
The studio has shipped before: Battle Derby, a live F2P mobile title, profitable with positive ROAS and an esports collab with Team Heretics. Grant-backed (Canarias Financia · ERDF), operating under the Canary Islands ZEC 4% corporate-tax regime.
Twenty years shipping games and film. Eleven running PromineoStudios on licensed IP for Disney, Warner and Marvel, with a PAX East showcase and a GDC finalist. Tutored executives for Oxford's Saïd Business School. Runs brand, community, Steam and the investor side, on an AI-agent stack he builds himself.
LinkedIn ↗
Nine years of Unity, senior in Photon Quantum multiplayer, full-time on the game. Built the whole RTS and the generation pipeline single-handed: 190+ units, six factions, the multipart builder and the in-game workshop. From prompt to combat-ready mesh in one click.
LinkedIn ↗Roblox proved players will build. Fortnite proved a free game prints money. AI made both programmable. GameGen sits where the three meet: a finished title, a creator layer, and a revenue share on every credit spent inside it. The category does not exist yet. That is the point.
The round funds three things:
Steam Next Fest in October, launch right behind it: servers, QA and the last mile of polish. The build is done; the round buys the runway to the door.
Seed credits for the first thousand creators, with the prompt exchange and the player market live at day one. Liquidity first, take-rate second.
Wishlist campaigns, creator collabs and the co-marketing push with the partner. Two people build the game; the round buys the megaphone.
The shape: ~€2M for an 18-month runway to launch and the numbers that unlock a Series A.
ENISA · CDTI Neotec · SODECAN / Canarias Financia · EIC / Horizon Europe, on top of the ZEC 4% corporate-tax regime. Extends runway at zero cap-table cost.
Gaming angels plus a European lead. Comparable seed rounds in this exact lane: Astrocade $12M, Series AI $7.9M (a16z Games), Bitmagic $4M (Supercell), Hidden Door $7M.
Next milestone → Series A: GameGen with real, paying player-creators.
The sponsorship deal on the table: structured as a recoupable advance against the studio's revenue share on in-game credit spend. They recover first; we earn after. Aligned by design.
Could a partner build this themselves? They could, and it would close the category: no studio builds on a platform that competes with its own clients. Their engineers reached out to us first. The seat this round buys is at that table.
Joshua learned the only winning move was not to play. In a new category, it is to play first.
The first step is small → a 30-minute walkthrough: the live build, the pipeline running, and the numbers behind this page.