r1n7aro/Locus

The open source Unity Dev Agent

What it solves

Locus for Unity is designed to reduce the tedious, repetitive work associated with game development in Unity. It provides an AI agent that can handle complex development workflows, from writing C# code to modifying Unity objects and assets, directly within the editor.

How it works

Locus operates as a standalone Rust + Tauri + Vue.js application that runs as an independent process. It uses a proprietary intermediate representation to help the agent read large scenes and assets, and leverages Roslyn for JIT-compilation and semantic analysis of C# code. This allows the agent to make edits and receive compiler-grade diagnostics without waiting for Unity's own compilation process. It also features a built-in C# hot reload system that applies changes instantly without requiring a domain reload, preserving the Play Mode state.

Who it’s for

Unity developers and creators who want to automate feature development, debugging, and asset management using an AI agent.

Highlights

  • In-editor operations: The agent can write C# code and modify Unity objects and assets.
  • Rider-grade static analysis: Provides live diagnostics, go-to-definition, and find-references using Roslyn.
  • C# Hot Reload: Applies code changes instantly without recompiling the assembly or running a domain reload.
  • Automated Knowledge System: Summarizes conversation requirements into design documents and preserves project understanding in long-term memory.
  • Modern Editor UI: Allows the agent to build Unity editor interfaces using Vue.js via the /view command.
  • Runtime Analysis: Uses C# state-machine tools to sample internal state through reflection for dynamic debugging.
  • Visual Version Control: Includes a graphical interface for semantic diff review and conflict resolution for Unity YAML files.

Related

  • Project
    hatayama/unity-cli-loop

    Unity CLI Loop is a Unity package + cross‑platform CLI that exposes 21 concise commands (skills) for compiling, testing, logging, hot‑reloading, UI manipulation, PlayMode input simulation, screenshots, etc. An LLM can invoke these commands to run a full development loop—compile → test → debug → verify UI—without a human opening the editor. Installation is a one‑liner for the CLI, a Unity‑package install, and a skill‑install step for the target LLM (Claude Code, Codex, …). The system talks to Unity via a Unix‑socket / named‑pipe, needs Unity 2022.3+, and is MIT‑licensed.

  • Project
    nutshellai-tech/mobius

    Mobius is a self-evolving Agent OS that connects teams, AI agents, and compute resources into a single workspace, allowing the system to rewrite its own code and UI based on user input.

  • Project
    Besty0728/Unity-Skills

    UnitySkills is a Unity 2022.3+ plugin that runs a local REST server exposing 805 editor “Skills”. AI assistants (Claude Code, Antigravity, Codex, Cursor, Kimi Code, OpenCode) can call these endpoints to create, modify, or delete assets, scenes, UI, networking, XR, and more. The system includes a three‑tier permission model (Approval/Auto/Bypass), dry‑run planning, structured JSONL audit logs, per‑task rollback, and batch transactions, all designed to keep AI‑driven edits safe and reversible.

  • Project
    ColeMurray/background-agents

    Open‑Inspect is an open‑source platform for AI‑driven background coding agents. It lets trusted team members launch AI sessions that run in full‑stack development sandboxes (Node, Python, git, browser automation) via a web UI, Slack, GitHub, Linear or webhooks. Sessions can span multiple repositories, support reusable “skills”, spawn parallel child sessions, and be triggered on schedules or events. The system runs on Cloudflare Workers (control plane) and various cloud sandbox backends, and supports multiple LLM providers (Claude, OpenAI, Grok, OpenCode).