# How do AI game makers work?

An AI game maker takes what you describe, sends it to a language model along with information about your project, and puts the code the model produces into a game running on a conventional engine. The interesting differences between tools are in that middle step: how much the model is told, and what it is allowed to change. Flockbay is free, AI included: free hosted AI every day, or your own Claude, ChatGPT, agent or API key. It runs on Windows, macOS and Linux, on the official Godot engine.

Source: https://flockbay.com/how-do-ai-game-makers-work/ · Updated 2026-10-24

## The basic loop

You describe something the game should do. The tool packages that description together with context about your project and sends it to a language model, which is running on a server somewhere rather than on your machine. The model returns code. The tool puts that code into the project and runs it.

You then play the result. If it is wrong you say what was wrong, and the loop runs again with that added to the context. Nothing more mysterious than that is happening.

This is why every tool in the category needs [an internet connection to build](/answers/does-flockbay-work-offline/). The model is on the network. It is also why "it made my game in thirty seconds" and "it took four attempts to get the door right" are both true statements about the same tool.

## The part that separates good tools from bad ones: context

A model cannot see your game. It sees whatever text the tool decides to send it. If the tool sends nothing but your sentence, the model is writing blind and will produce plausible code that references things which do not exist in your project.

So the engineering work in an AI game maker is almost entirely about what goes into that context. What nodes exist. What the scene contains. Where the terrain is and what the weather is doing. What happened last time code ran. Every one of those is a thing the model no longer has to invent.

Removing the guessing is what [turns bad output into a working game](/blog/why-your-first-ai-generated-game-is-bad-and-what-to-do/), and that is work on the tool: what it sends the model, and what it lets the model run and check.

## What the tool does with the model's answer

In the simplest tools, it writes a file and hopes. In better ones, the code goes into a live project and the tool watches what happens — whether it compiled, whether it errored at runtime, what the game did — and feeds that back so the next attempt is informed rather than a fresh guess.

The other thing the tool controls is scope: what the model is allowed to touch. Some let it rewrite anything, which is fast and makes changes you did not ask for. Some bound it to code and leave the world alone. That choice is a product decision and it is worth knowing which one you are using.

None of this involves AI at runtime. In almost every tool the model is a building tool, and the game you end up with is [ordinary code that runs on its own](/answers/can-i-edit-the-code-myself/) with no network and no bill.

## How Flockbay does it

Flockbay is a desktop app for Windows, macOS and Linux with Godot bundled inside it. You describe the game, and Flockbay agrees a brief with you first: what gets built first, what comes after, what is not in this version. Once you approve it, Flockbay builds the world, writes the GDScript, runs the game itself to check it, and then you press Play — the game runs locally on that copy of Godot.

The context problem above is what the capability layer exists for. It gives the model a world it can address by name rather than a folder of scripts to reconstruct from. Which names depends on whether the project is 2D or 3D, which the app preselects from your first message. In a 3D world the vocabulary is terrain, water, weather, the player controller and the props standing in the scene. In a flat one it is the tile ground, the band of scenery on the horizon, the player controller the camera follows and the sprites in the scene. Same layer and the same chat; different nouns.

The model is bounded to what it can call. It builds your world only through the world tools, and the result is real nodes in the scene tree. It checks its work the same way: a headless run or a window, input events, stdout and the debugger, a viewport capture. And there is no path where Flockbay asks for something and the app quietly does something else instead — either Flockbay does it with those tools, or you press the same tools yourself, as buttons around the Godot editor's view: [the manual controls](/plugins/).

## What this means for you in practice

Building needs an internet connection because the model, asset generation and world data all come over the network. Playing what you built does not, because the finished game is ordinary Godot code.

Iteration is fast because the game runs on your machine rather than in [a render queue somewhere](/cloud/). And the project is a normal Godot project on your disk, so nothing about the mechanism above locks the result inside the tool that made it.

## Questions

### Does the AI run on my computer?

No. The model runs on a server and reaches it over the network, which is why building needs a connection. The game itself runs locally.

### Is the AI inside my finished game?

No. It is a building tool. What ships is ordinary Godot code with no model in it.

### Why does it sometimes get things wrong?

Usually because it was missing context about your project. That is the problem the capability layer addresses, and describing what you saw is how the next attempt gets it.

### Can I use my own AI agent instead?

Yes. Flockbay ships one MCP server, so an agent that speaks MCP can drive it. Connecting your own agent is free. You pay your agent's provider for the model; Flockbay charges nothing for the connection.

### What language does it write?

GDScript, into your Godot project, where you can read it if you want to.

Download Flockbay free: Free to download for Windows, macOS and Linux. https://flockbay.com/

Related: [What is an AI game maker](/what-is-an-ai-game-maker/), [What is MCP for game development](/what-is-mcp-for-game-development/), [Does AI game making actually work](/does-ai-game-making-actually-work/), [What is GDScript](/what-is-gdscript/)