Code Grid

About This Game

Hidden away in your father's study is a story you grew up believing he'd made up—the “Grid Computing Machine.”

A machine long since fallen silent. Stacks of yellowed manuscripts. Notes from four generations filling the margins, questioning and contradicting one another. Documents, letters, recordings, and experimental reports preserve scattered traces of its past.

As you delve deeper into its puzzles, you'll gradually uncover the people who devoted themselves to the machine—and why it was eventually forgotten.

Some explorations may ultimately prove impractical.

That doesn't mean they were without value.

Build Your Program on the Grid

You don't need to write a single line of traditional code.

Place instructions directly on the grid, plan the machine's path, and set it in motion. It travels across the grid, executing each instruction it encounters—reading numbers, changing values, altering direction, and ultimately producing the correct output.

Your goal is simple:

Build a program that passes every test.

Place the Instructions. Set It in Motion.

Every cell can become part of your program.

Place instructions, redirect the machine, and guide it toward what you want it to execute next. When everything is ready, start the machine and watch it follow the route you've designed.

Here, a program isn't just a sequence of commands running from top to bottom. Where instructions are placed, the order in which the machine encounters them, and the path itself are all part of the program.

Simple Instructions. Complex Programs.

The basic rules of CodeGrid are simple.

But once loops, conditions, state, and data processing begin to interact, the same small set of instructions can produce entirely different behaviors.

As your program runs, numbers change, paths branch, and data flows through the machine. A program built from only a handful of instructions can gradually become a complex, finely tuned machine.

The grid is finite. The possibilities are not.

Solving It Is Just the Beginning

• Three-Dimensional Scoring — Every puzzle evaluates your program across three dimensions: instruction count, execution steps, and cost. Passing the puzzle is only the first step. The real challenge is making your solution leaner and more efficient.

• Global Leaderboards — Push for fewer instructions and lower costs, and compare your solutions with players around the world.

• Self-Modifying Code — Let a program rewrite itself while it runs. The two most mysterious cells on the grid will completely change the way you think about building programs.

A Conversation Across a Century

From Victorian-era blueprints to a prototype built in wartime, and finally to code written by an old man teaching himself to program late in life—four generations took up the same unfinished problem, leaving questions for one another in the margins of their manuscripts.

None of them lived to see the answer.

But every manuscript is still waiting for you to run it.