Cookie Clicker: my own tools

Cookie Clicker is a game where you click a cookie. That's it, and that's the joke it opens with. Then you buy a cursor that clicks for you, a grandma who bakes, a farm, a mine, a portal to another dimension, and three years later you find yourself with a save you never abandoned.

I played it in secondary school, like a lot of people back then. I came back to it much later, and ended up buying the Steam version. Along the way I built two tools out of it, three years apart, which say a fair amount about my relationship with the game and about what I could do at each point.

A game with far more depth than it lets on

Under the joke sits a dense set of mechanics. Dozens of buildings to level up, upgrades, achievements, and above all ascension: you sacrifice your run to start over with a permanent bonus. On top of that come genuine minigames, including a garden where you crossbreed plants, a stock market, a pantheon and a spell grimoire.

It's also a game that speaks to developers. It runs in the browser, everything is JavaScript, and the game makes no secret of it: it eventually sells you a building called the "Javascript console". It exposes a real modding interface, which the community has used to build entire add-ons, and a huge wiki catalogues every plant, every probability, every crossbreeding recipe. Part of the fun is reading that documentation as much as playing.

The period when I cheated

That transparency has a consequence: cheating is trivial. The game state lives in a JavaScript object reachable from the browser console, and one line is enough to hand yourself a billion billion cookies or unlock whatever you like. Entire sites list those commands.

That's how I started. Console open, commands pasted, every possible tool alongside. The game isn't fooled: in that case it awards a hidden achievement whose name sums it up nicely, "Cheated cookies taste awful".

After a while it wore thin. A run where anything can be had in one line isn't worth much. So I dropped that save and started a clean one, on 3 October 2022, never touching the console again. It's the one I still play.

My first extension, three days later

On 6 October 2022, three days after starting that run, I wrote my first browser extension.

Golden cookies work simply: they appear at random intervals, stay a few seconds, and grant a large bonus if you click them. So you have to keep an eye on the screen constantly, which is the opposite of what you want from a game running in the background.

The extension is ten lines. It repeatedly looks for elements carrying the golden cookie class and clicks them. Nothing more, and honestly nothing clever: the loop runs at the browser's maximum rate, without the slightest delay, which nobody with a bit of experience would write. It only worked on Firefox and had to be loaded by hand each session, as a temporary add-on.

It still belonged to my "help myself with everything" period, and I stopped using it along with the rest. It remains my first encounter with the extension format, the one I'd come back to much later on far more ambitious projects.

What my save says

My current run
Started on	3 October 2022, about 1,390 days
Ascensions	13
Buildings owned	15,500
Cookies baked in total	7.75e+61
Golden cookies clicked	5,182 all time, 1 in the last 108 days
Garden	level 11, 34 seeds out of 34

That golden cookie line is the clearest trace of the change. Across the whole run, over five thousand clicks; across the last hundred days, a single one. I no longer watch the screen and I no longer automate anything, I just let it run.

The garden, the minigame that deserved a tool

The garden is the trickiest minigame. You lay plants out on a grid, and two neighbouring plants can spawn a third, with a probability that depends on the species, the soil type and maturity. Some species contaminate their neighbours. Unlocking all thirty-four seeds means chaining crossbreeds in the right order, with the right layouts, waiting through cycles of several minutes.

Doing that from memory is tedious, and juggling the wiki and the game is just as bad. So in 2025 I built an interactive guide: the full list of seeds, each one's crossbreeding recipe, the layouts to use, the properties of each soil, the contamination traps, and a tracker for what's left to unlock.

Unlike the extension, this tool doesn't play for me. It only gathers and presents information that already exists, the way a better-organised wiki would. That's why I used it without second thoughts, and it did help me complete my thirty-four seeds.

A one-day project, owned as such

The site was written in a single day: most of the commits are dated 1 July 2025. The stack is deliberately minimal, Next.js, Tailwind and a few UI primitives, with no state management library and no data-fetching client.

The home page actually advertises ten tools, including a production calculator, a building optimiser and a stock market helper. Only one really exists, the garden one, the only one I actually needed. The other nine stayed as titles.

That's roughly the lesson of the project: sizing the effort to what's at stake. A tool for a browser game needs no layered architecture, no integration tests and no deployment pipeline. Recognising that a project is small, and treating it as such, saves you three evenings spent building scaffolding for a one-room house.

Three years between the two

Ten lines in a file in 2022, a complete application in a day in 2025: the same game served as the testing ground for both, at very different points in my learning.

The 2022 extension is also what gave me the reflexes I reused much later on a far more complex project, a bot for Microsoft Rewards, where it was no longer about clicking a cookie but about convincing a site that nobody was automating anything.

As for the run, it's still going, with no commands in the console.