The photo becomes documentation
The starting point for MakerMaker was one question: what if handing a photo to an AI produced something genuinely creative — not a caption, not alt text, but a piece of weird writing?
Three hand-built generators ship with the app. The first takes any photo (or text you type) and produces a fictional corporate operations manual — deliberately bureaucratic, the kind that reads like it survived three reorgs: version numbers, revision dates, department headers. The second turns photos into imaginary product spec sheets. The third writes breaking news headlines around whatever you feed it.
Three generators is enough for a few minutes. It's also a ceiling that comes fast.
AI inventing its own generators
The feature that made the project worth shipping: a button that tells the AI to design and then run an entirely new type of generator. It picks subject matter, genre, tone, format, and color palette — roughly 95 million combinations across those axes.
Gemini 2.5 Flash runs on Firebase Functions in asia-northeast1. The AI decides what kind of thing to create before it creates it. Results range from "Showa-era company newsletter" to "academic abstract" to genres we didn't anticipate. About 30% of outputs are unremarkable — that's a real number and we're not hiding it. The other 70% are worth the tap.
What we chose not to build
Photos are not stored server-side. Processing happens on Firebase Functions and data is discarded immediately. Non-negotiable from the start.
The UI is Japanese only. We traded reach for focus — understand one market before spreading thin. It limits the audience for now.
Download
Free to download. Generation costs points (10pt for ¥100), with a ¥480/month premium pass for heavier use. Requires iOS 17+.
Top comments (0)