I built a mobile dungeon defense game alone, with Unity and AI tools, and shipped it on the App Store 39 days after creating an empty project. This series follows that run day by day, then keeps going after launch with real revenue numbers, using only the commit history and screenshots taken at the time.
The main text is enough. It is written so you can follow how the game came together without any coding background.
Open the Dev notes in each post for structure, code decisions and the reasoning behind them.
EP.142026-10-23 ~ 11-02
For the first week after launch, in-app revenue was zero. The first purchase came on day 8, and every one of the three people who paid bought the pack that lets you skip ads. Ad revenue: $3.23.
Read ↗
EP.132026-09-14 ~ 10-23 · 407 commits
407 commits between creating an empty project on Sep 14 and launching on the App Store on Oct 23. Where I used AI tools, where they got stuck, and what I had to hold onto myself.
Read ↗
EP.122026-10-21 ~ 10-25 · 50 commits
I submitted on the 21st and pressed the release button on the morning of the 23rd. Ads didn't show that day, so I made 1.0.1, and the day after I moved every line of text in the game into English.
Read ↗
EP.112026-10-13 ~ 10-21 · 130 commits
In nine days I shipped Android internal test builds with version codes 2 through 13 and five TestFlight builds for iOS. In between I re-recorded the voices, rebuilt the dungeon screen, made the screenshots, and hit "Submit for Review" at 4:58 PM on October 21.
Read ↗
EP.102026-10-01 ~ 10-16
I added nine rewarded ads, built a shop and tore it down again, and chose a payment server. Two weeks that ended with my first test purchase and my first refund.
Read ↗
EP.092026-10-02 ~ 10-14 · 14 commits
Saving gave gold back, a full mailbox ate gear, and one stale backup locked the shop. Two weeks in early October spent almost entirely on save data.
Read ↗
EP.082026-10-07 ~ 10-11 · 62 commits
The adventurers standing next to my monster staff looked like they came from another game. On October 7 I redrew all 54 of them in one day, then went on to the gear icons, effects and attack animations.
Read ↗
EP.072026-09-29 ~ 09-30, 10-04 · 25 commits
Once all ten story chapters were in, the obvious question was "what do you do after the ending?" I added four guildmasters, T6 gear and 118 hard mode stages, then ran simulated players for 84 days to tune the difficulty.
Read ↗
EP.062026-09-24 ~ 09-29 · 35 commits
I added Chapters 4 and 5, then redrew the 19 cutscenes I'd faked with composites using PixelLab that same night. Three days later I wrote Chapters 6 to 10, 60 more stages, through to the ending. The art had to wait another eleven days.
Read ↗
EP.052026-09-22 ~ 09-23 · 37 commits
On the morning of the 22nd I took the roster from 50 monsters to 180. Then I spent two days making battles feel like something: hit stop, knockback, a slow-motion finisher, layered sound effects and two battle themes.
Read ↗
EP.042026-09-18 ~ 09-21 · 59 commits
On the 18th, 40 commits went into a tutorial, a story script and an app icon, and the game's progression changed shape. On the 21st I played it on my phone and fixed most of what I wrote down in one day.
Read ↗
EP.032026-09-17 · 8 commits
An eight-commit day. Monsters learned to level up and awaken, hero loot became a real gear system, and all 24 battle sprites got redrawn. At 4:18 p.m. the game got its name.
Read ↗
EP.022026-09-15 ~ 09-16 · 15 commits
The day-one prototype lasted two days. In its place I wrote 18 design docs, rebuilt the combat engine, made a result screen and settled on pixel art.
Read ↗
EP.012026-09-14 · 10 commits
At 4:30 p.m. I created an empty project. By 5:30 there was a dungeon where you place monsters and they fight off heroes. This is everything that happened that day, up to late at night.
Read ↗
NOW AVAILABLE
Heroes are intruders, monsters are staff. Save Dungeon No. 7 from being shut down.
Available on the Korean App Store (iPhone/iPad), with English in Settings. Android is not out yet.
