News, updates, and dev logs from Inselnova, a browser-based island strategy game.
Founders World, week in review: patrol boats kept failing to return, an unknown raider kept its name line perfectly empty, and the Harbor Board answered with counting, recounting, and more liberated Black Tide shores than anyone cares to initial. roxx and Miko Misan measured Matte1cat by force while Vexed Flamingo measured the marketplace by weight.
Six months and just over a thousand governors into Inselnova, a browser strategy game: who actually stayed, how search quietly became a steady source of players while I wasn't marketing, and the three-stage plan from here.
Founders World week in review: Flowkap and Aye-for-an-Aye struck Medium Island against the Black Tide, each returning with a blackened idol as other Tide shores were recorded as liberated, some without a name. roxx ran a long, tidy raid-circuit through BigChris' islands without losses. Miko Misan broke into the top ten on a surge of raids, while Lusty Longboat raised a fresh charter at…
Send fleets from several islands at one target so they land together, then commit yours to an alliance strike that sails on an officer's order.
Founders World, week in review: roxx drives clean raids through BigChris' dormant outer islands, then orders an independent audit at SplashFish Nook. The Black Tide loses ground at Medium Island, Black Tide Chills, and Black Tide Pall, while BoomStinger Landing sees fresh repulses. Hidden raises a new flag at MapSquid Quay as Luminous Reveller's name briefly touches the top ten on a posted board.
Set a shipment between two islands once and a crew loops it for you. Research becomes a climbing tree, and your world prints its own daily paper.
After a year of AI-assisted development, I've lost the thing I was proudest of: being able to debug my own code in my head. On the trade behind it, and what it quietly costs.
Most people run AI one agent at a time, which wastes their time. I pushed to four and the cognitive switching burned me out. Here's the system that made running multiple AI coding agents work: every session starts from a ticket, a small /now-what skill catches me up, and the work is split across separate editors.
Saturday review for Founders World: Black Tide-held ground was cleared again at Medium Island as Flowkap and Aye-for-an-Aye each forced a victory and brought home the same curious spoil. Aye-for-an-Aye also climbed to #8 on the realm board after new colonization at CrabPickle Watch and heavy plunder. Windy Walrus remains the realm's largest presence, while smaller courts like kelti, roxx, House Of…
Capture or liberate any rival colony regardless of login, survive punching up, and watch the sea come alive.
I build Inselnova test-first: ask AI for a failing test, then let it code until green. Better code, but 5,000 AI-written tests get slow and brittle.
After months on the mid-game, players who reach the late game keep asking what they're playing toward. The old browser strategy games answered that with a server reset. I'm building a real end-game instead: a second world you choose to sail into.
Claim a stretch of open ocean and let its residents vote it yours, raise motions in the new Alliance Council, and feel a rebalanced plunder game.
A tour of the Claude Code skills I use to build a browser strategy game, in the order I actually run them: scope a ticket, balance the economy, check the cascade, write a failing test, ship the version.
How to write and chain AI coding skills so a coding agent works your way: less typing, fewer bugs, and repeatable results instead of guesswork.
How a project shifts from feature factory to steady, tested releases once real users arrive, shown in a browser strategy game with real numbers.
Raise a Call for Aid and your alliance sails to defend you. Alliances now climb six ranks, and the new Tavern deals games of nerve.
I ask AI to investigate my game's database instead of building dashboards. It changed how I think the end of SaaS plays out.
Three new ways to fight at sea: raise a patrol to guard your waters, blockade enemy convoys, and read raiders off your Watch Tower.
I'd never built a game, so I started with a loop, not a plan. Five onboarding rebuilds, each one a data-driven experiment rather than a step on a roadmap.
A new stealth Plunder raid that slips into enemy harbours, plus alliances that earn XP from quests, raids, and just showing up.
Three months of live data on the players who quit my browser strategy game mid-game. The most invested ones don't fade out, they stop all at once.
Codex writes better code, but after months running both I moved everything back to Claude Code. The difference is iteration size, not quality.
Your harbour now caps how many ships an island can hold, plundering counts double, and Tideline Finds wash ashore as a sealed sea-crate.
Dread Lord ran the longest win streak the Founders Archipelago has recorded, forty-five straight, then stopped coming to the water and, by his own word, retired undefeated. Windy Walrus now sits first with two hundred and six islands. The Black Tide has been thrown back two hundred and sixty-eight times in a fortnight and is still arriving.
Why I've been shipping features into my multiplayer browser strategy game out of fear, what my retention numbers do and don't tell me, and why the game doesn't know what it is yet.
There's talk of an indie golden age. The barrier to making an old-school browser strategy game has fallen away, so they're everywhere. That was never the hard part. The hard part is deciding what to build.
My AI loop workflow finds, fixes, and proves bugs in my own systems. Here's how it works, what it needs to be safe, and the work I won't hand it.
A wave of new browser games look like the old ones, and people call them clones. It's a genre, the same way Minecraft clones became survival-crafting.
Let an island's food stores hit empty and the garrison starts to die, Colony rank and up, and blessings now count when the fleet lands.
I built an AI marketing mentor by priming three chatbots and letting Claude Code distil the output into a reusable skill. The trick is priming.
A new alliance event where the island sinks and drowns your troops, goods that rush any job, and a safe way to ship lumen.
My king-of-the-hill event for a browser strategy game broke twice. One alliance walled off the islands in thirty minutes. What failed, and the fix.
I sped up my browser strategy game to ask less of players. The survey said I'd done the opposite. Faster meant more check-ins, not fewer.
Inselnova now plays in German, and a new war god, Korvath, opens up the first time you raze an enemy island.
Most strategy games want your whole evening. Inselnova wants about five minutes, twice a day, in your browser. The fun part is getting a few coworkers in, so there's a quiet war running under the workday.
Save a send as a reusable fleet template and reload it to pre-fill the dispatch in a tap, plus a rebuilt Realm overview.
A once-a-day Morning Report opens after time away, with overnight tallies, what Brann found on the shore, and a Days of Rule streak.
Raise a wealth statue at Main House 15, feed it coin to climb Bronze, Silver, and Gold, and rise up the new Wealthiest leaderboard.
Most persistent strategy games eventually wipe the world and start over. Inselnova never does. An NPC faction called the Black Tide takes over abandoned islands and attacks their neighbours, so the world cleans itself up instead of resetting.
Sinking ships leave salvageable flotsam, a new Community tab, redesigned fleet cards, and a batch of economy balance changes.
Hold contests get a live Ledger showing every attack, settlement, and capture as it happens, plus a new Events tab and Nyssa's Trade Wind rite.
The Black Tide now raids your lumen and stashes it on its own islands. Watch the hoard from your tower, raid it back, and bring home a Blackened Idol.
Close every island for up to 21 days so no one can raid you while you're gone, and run your alliance from a shared Operations war room.
Inselnova is my first game, built solo with AI, three months live with real players. The AI is fast, and just as fast at shipping a mistake. The real work was the harness: a set of skills and gates every change has to pass before it reaches live.
Alliances fight over contested islands, you can bless allies and your other islands, a lost realm gets a comeback start, and a new pirate hunt lands.
Workshops let your island pick a trade and craft its own goods, and seizing a Forsaken island now costs the same Expansion Stress as empty land.
The early game is the gate to everything else, and too many new players weren't getting through it. Here's what five rebuilds of the first session in my browser strategy game taught me about where people stall and what actually keeps them.
Inselnova started as a copy of Inselkampf and has gone its own direction. A handful of additions came in: a bazaar, coins, council affairs, the Black Tide, a church with gods and a new resource called lumen, and a guildhall. Each one came from a real problem in the early or mid game. Here's the story behind each, and the rules that have crystallised along the way.
I built Inselnova because I wanted to build something. The audience that showed up is mostly returning Inselkampf players in their 40s. The early game feels slower for them than the original ever did. Here's what's pushing the next features in a different direction.