Building Movesia: Why We Created an AI Assistant for Unity

The Problem We Set Out to Solve

Game development in Unity is powerful but complex. A typical Unity project involves hundreds of GameObjects, dozens of custom scripts, and a scene hierarchy that can grow to thousands of nodes. Developers spend significant time navigating the Editor, searching for components, and writing boilerplate code that connects systems together.

We noticed that most AI coding assistants treat Unity projects like any other codebase — they can help you write C# scripts, but they have no understanding of your actual scene. They don't know what GameObjects exist, which components are attached, or how your prefabs are structured. The gap between "AI that writes code" and "AI that understands your game" was the starting point for Movesia.

But the problem goes deeper than tooling. Indie developers and small teams often have the creative vision for a great game but lack the bandwidth to handle every aspect of development — from asset management to scene composition to debugging obscure runtime errors. We wanted to build something that doesn't just assist with code, but acts as a true development partner that can help a solo developer create a full, professional-quality game.

Why a Desktop App?

Unity developers already split their time between the Unity Editor and various tools. Rather than building a plugin locked into a single code editor, we wanted to give developers a dedicated companion app that works alongside Unity regardless of which editor they use for code. A standalone desktop app means no dependency on any particular IDE — just launch Movesia, connect to Unity, and start working.

By building Movesia as a desktop app, we could deliver a focused, purpose-built experience. No context switching between panels, no fighting for screen real estate with code tabs — just open Movesia and start talking to an agent that already knows your project.

The Architecture

At its core, Movesia is a LangGraph agent powered by Claude that runs inside the desktop app. The agent has access to six specialized tools for interacting with Unity:

• Query Tool — Search and filter GameObjects by name, tag, layer, or component type
• Hierarchy Tool — Browse the scene tree with filesystem-style navigation
• Component Tool — Read and modify component properties on any GameObject
• Prefab Tool — Instantiate, modify, and manage prefab instances
• Scene Tool — Load, save, and switch between scenes
• Refresh Tool — Force-sync the agent's understanding of the current scene state

Each tool communicates with a lightweight C# package running inside the Unity Editor. The package exposes a WebSocket server that accepts JSON-RPC commands and returns structured data. This design keeps the Unity-side code minimal and the intelligence on the app side, where it's easier to iterate and update.

What Makes It Different

The key insight behind Movesia is that an AI agent needs structured access to game data, not just source code. When you ask Movesia to "add a health bar to the player," it doesn't just generate a script — it finds the Player GameObject in your hierarchy, checks what Canvas exists in the scene, creates the UI elements, attaches the right components, and wires up the references.

This level of integration is only possible because the agent maintains a live connection to the Editor. It can inspect the scene, take actions, and verify the results — all in a single conversation turn.

The Movesia Asset Store

One of the biggest barriers for indie developers is access to professional-quality assets. Buying asset packs, importing them, organizing them in your project, and placing them correctly in scenes takes time and money. Movesia changes this equation with the Movesia Asset Store — a curated library of high-quality assets that the agent can pull directly into your Unity project for free.

When you tell the agent "I need a medieval castle entrance for my scene," it doesn't just point you to a download link. It browses the Asset Store, selects the right asset, imports it into your project, and places it in your scene at the position you want. You can then ask the agent to adjust placement, scale, rotation, or swap it for a different variant — all through conversation.

This is especially powerful for solo developers and small teams. Instead of spending hours searching for assets, importing packages, and manually positioning objects, you describe what you want and the agent handles the rest. The result is that a single developer can build scenes that look like they were assembled by a full team.

Error Detection and Auto-Fixing

Writing code, wiring up components, and managing a growing scene inevitably leads to errors — null references, missing components, broken prefab connections, script compilation failures. Traditionally, you'd see an error in the Console, track down the source, and fix it manually.

Movesia's agent can detect these errors as they happen and fix them on the spot. When the agent writes a script or modifies a scene, it checks for errors in the Unity Console and resolves them immediately. If a serialized reference is missing, the agent finds the correct object and assigns it. If a script has a compilation error, the agent reads the error message, identifies the issue, and patches the code — all within the same conversation turn.

This tight feedback loop means fewer interruptions and less time spent debugging. For indie developers working without a QA team, having an agent that catches and fixes problems in real time is a significant advantage.

What's Next

We're actively developing Movesia with a focus on three areas: growing the Movesia Asset Store with more professional-quality assets available for free, expanding the agent's capabilities to cover more Unity Editor operations, and building a community of Unity developers who shape the product through feedback.

Our vision is simple: an indie developer with a great idea should be able to build a full, polished game with Movesia as their development partner. From scaffolding scenes with professional assets to writing code, detecting errors, and fixing them automatically — every part of the workflow should feel like you have a capable team behind you, even if it's just you and the agent.

If you're building games in Unity and want an AI assistant that truly understands your project, we'd love for you to try Movesia and tell us what you think.