Loading page…
A former game artist who never writes a line of code himself is building a Unity card roguelike entirely through Claude Code, sustained for 8 months by a workflow of "search-engine-style prompting + subagent review"
no revenue yet (demo in development)
He treats Claude Code as a search engine, not magic — card roguelikes have been built a million times over in the training set, so the point isn't one-shotting it, it's a workflow comfortable enough that you don't quit halfway.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
The protagonist worked as an artist at a game studio. He can't write code and doesn't intend to, yet he wanted to independently build a card roguelike called Mythmatic. That sounds like the classic "AI builds a game in one click" gimmick, but what he actually delivered isn't a shipped product — the game still isn't released, has zero revenue, and is still being cut down toward a playable demo. What he delivered is a workflow that kept him from quitting for 8 months. That's precisely why it's worth including. The highest mortality rate in indie games isn't "couldn't build it," it's "gave up midway." Most people doing an AI side project are excited for three days and burned out by week three, because feedback is too slow and the isolation is too heavy. His fix was to turn the whole development process into a YouTube devlog, converting "grinding behind closed doors" into "building in the open with an audience watching," using an external feedback loop to keep himself alive. His framing of AI is equally level-headed: Claude Code isn't magic, it's a search engine. Card roguelikes have been built a million times over in the training set, so the model already "knows" how to do it. The point isn't hoping it one-shots the right code, it's using the right prompts to extract the solution already in its head, then polishing it repeatedly with a comfortable process. Get that mindset straight and every concrete technique below stands on solid ground.
There are no customers yet, and that is deliberate — his position is that marketing has one real precondition, a demo, so until then the only channel he runs is the YouTube devlog. It is already doing acquisition work anyway.
The quick-reference card — acquisition channels, replication playbook, and risk map — is members-only.
Join the Unicorn Club — $5/moCancel anytime, in one clickPaste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: The video provides no revenue figures at all; the game hasn't launched and is still in the cut-features, build-a-playable-demo stage. Included as a methodology reference. All revenue figures are founder-reported or estimated by the video creator — unaudited.
Nothing has been charged yet, and in this category the first dollar has a fixed shape. The demo is free, because the demo is the marketing, and money starts the day the paid build goes live in the ten-to-twenty-dollar band where indie deckbuilders sit — not at whatever the months of work would justify.
The compounding here is in the workflow, not yet in the market. What turned was finding a process comfortable enough that quitting stopped being the likely outcome, and each piece makes the next feature cheaper than the last.
Eight months in with nothing shipped, he calls it developing into the void, and he is direct about what makes that bearable.
The rest of this breakdown — 10 more sections — is in the Unicorn Club, with every section of all 1,335 ideas: first customers, first dollar, the flywheel, and the silent stretch.
Join the Unicorn Club — $5/moCancel anytime, in one click