A game design document is not a game. I know that. You know that. I have spent the better part of the last few weeks with a GDD that I am genuinely proud of, and zero lines of code to show for it. That changes this week.
But before the build begins, I want to write down what exactly we are building and why — because in six months, when I am knee-deep in Godot scenes and bug reports, I will need to remember what this was supposed to feel like.
The Mechanic, Plainly Stated
Mosaic is a puzzle game where you drag image pieces into their correct positions on a grid before a timer runs out.
That is it. That is the core. A shattered image, a grid of empty slots, and a clock that does not care how you are doing.
The pieces sit in a bank on the side of the screen. You drag one, hover over the board, and drop it near its correct slot — it snaps into place if you are close enough. Get them all placed in time, and you move on. Run out of time, and something happens. Something bad. You have been warned.
What the image shows is the other half of the mechanic. You do not know what you are assembling until it comes together. Most images are fine. Some are not. Completing a forbidden image has consequences of its own — the game punishes curiosity as much as failure.
Why This Is Not Just a Puzzle Game
The mechanic earns the horror. That was the design rule from the start.
I did not want to bolt horror onto a puzzle as a coat of paint — a jumpscare that plays regardless of what you did or did not do. In Mosaic, the jumpscare is a consequence. You either ran out of time, or you finished something you should not have. The dread comes from the fact that both failure states are your fault. You assembled it. Piece by piece. You did this.
That is the feeling the game is built around. Not gore. Not shock. The slow, compounding unease of a clock ticking while you try to figure out what you are looking at.
The Structure Around the Puzzle
Every run is a node map in the style of Slay the Spire — three lanes, three nodes deep, with a boss at the end. Each node is a puzzle encounter. Minion nodes are faster, smaller grids, and reward gold on completion. The boss nodes are harder, longer, and carry unique rule modifiers that change how the puzzle behaves.
The five bosses each break a rule you thought was fixed. One makes placed pieces drift and un-snap after a few seconds. Another hides thirty percent of the pieces until you hover directly over them. The modifiers are not gimmicks — they are the boss fight. Learning how to handle them is the skill progression.
Between runs, a persistent meta-layer builds out. Gold carries over across deaths. Cards — thirty seconds of bonus time here, a piece highlight ability there — are collected, upgraded, and slotted before each run. You equip three per run from a growing collection. You get stronger across sessions even when you lose a run.
Where We Are Right Now
The GDD is complete and locked at version 1.2. The studio brand is live. The itch.io page exists. The cover art is placeholder, but the foundation is set.
What does not exist yet is the game itself.
Phase 1 starts this week. The goal is a working core puzzle mechanic in Godot 4.6 — drag and drop, grid snapping, a functional timer, and a basic consequence state when that timer hits zero. No map. No cards. No bosses. Just the thing the entire game rests on, proven to actually work.
If the puzzle is not fun at its most basic level, nothing else matters. So that is what we are testing first.
What Comes After That
Once the core mechanic holds up, the map system comes next, then card integration, then the first boss. The full pipeline is planned. The scope is defined. I am not building a vague concept — I am building from a document with a risk register and a milestone table.
That said, plans change. If something does not feel right in the build, it will change. That is what these devlogs are for — to document what actually happened, not just what was supposed to.
Follow Along
If you want to watch this get built from scratch, follow the project on itch.io. That is where the early demo will land when there is something worth playing.
The next devlog will have actual screenshots. Or it will have an honest account of why that did not happen. Either way, it will be the truth.