How do I use Flockbay with my own AI agent?

A game project is normally opaque to a language model. It is scenes, binary resources, an editor that holds state, and a thousand conventions that live in a person's head. An agent asked to change something has to infer most of it and it gets things wrong in expensive ways.

Flockbay exists to close that gap. It is a free AI game maker built on Godot: a desktop app, on Windows first, that bundles the official Godot 4.7.2 engine, keeps the project in a real Godot layout, and exposes the work as things an agent can actually do rather than pixels it has to guess at. That is the layer, and it is why the code an agent writes here lands correctly more often than the same model writing into a bare project.

Flockbay's own AI is in the app. It writes GDScript and scenes into your project, shapes the world through the capability layer (terrain, water, sky, weather, the player controller and props), and playtests the game to check its work: it runs it, reads stdout and errors, captures viewports and injects input events.

You judge the result by playing it, and you can take over by hand at any point. The builder is a set of buttons around the Godot editor's own 3D and 2D view, with Godot's undo, so you and the AI use the same controls.

Every account gets free AI each day on the hosted model, called Demo. When the free usage for the day runs out, it comes back the next day. You can also run it on your own Claude plan or ChatGPT sign-in, or on your own API key, connected once in Connectors. Flockbay has no paid plans.

More on this: Claude for Godot: how to build a Godot game with Claude Code, step by step

If you already work in Claude Code, Codex or Cursor, point it at Flockbay over MCP. It is one server, registered as flockbay-stage, and the app writes the line to paste: Connectors, then Another agent, then Set up.

Your agent gets the same tools as Flockbay's own AI: the world capabilities, editing the game through the open Godot editor with undo, and running, watching and playing the game. Connect your own agent has the setup for each client.

It is also the version that is honest about what an integration is. You are not getting a different Flockbay; you are getting the same capabilities addressed by a different client.

More on this: Claude or ChatGPT for Godot

The game runs locally. Godot ships inside the app and pressing play starts the game on your machine, at your framerate. There is no streaming, no remote render and no browser preview.

Building needs an internet connection. The agent, asset generation and world data all come over the network, so this is not an offline tool and there is no degraded offline mode. Playing what you built is a different matter and needs nothing.

Free to download, and the builder itself is free. Every account gets free AI each day on the hosted model, called Demo. You can also build on your Claude plan (Pro or Max) or ChatGPT account (any plan, Free included) in the app, your own agent over MCP, or an API key, and you pay that company, never Flockbay. Pictures and sound run on the free AI. A 3D model uses your own Tripo key. Flockbay has no paid plans.

External agent access over MCP is free, with a free Flockbay account. There is no seat pricing and no enterprise tier; this is a desktop app one person installs.

More on this: Flockbay vs Summer Engine: official Godot, or an engine built on it

The interesting result is not that a model can write GDScript. It is how much less it has to invent when the surrounding project is arranged for it. Every ambiguity you remove — where a thing lives, what a scene is called, how a system is expected to be wired — is a decision the model no longer has to guess and get wrong.

That has a practical consequence. Reliability comes from the layer, not from reaching for a larger model, and a layer that makes a small model reliable is worth more than a prompt that makes a large one occasionally brilliant.

It is also why quality complaints here turn into work on the capability rather than a model upgrade. If the same request keeps producing the same wrong shape, the honest fix is upstream of the model.

More on this: How to make games in Codex: the Codex CLI, the ChatGPT app, or your ChatGPT plan inside Flockbay