About this project

HybridCLR is a full-platform native C# hot-update solution for Unity. By extending the il2cpp runtime and introducing an efficient register-based interpreter, it transforms the pure AOT runtime into an AOT+Interpreter hybrid mode, enabling dynamic assembly loading at the underlying level. It supports all platforms that il2cpp supports, including Android, iOS, consoles, and WebGL. Core features include: - Nearly complete implementation of the ECMA-335 specification, with only a very small number of features unsupported - Zero learning cost: hot-update code works seamlessly with AOT code, supporting generics, reflection, inheritance, etc., without requiring special code or code generation - Full support for multithreading (volatile, ThreadStatic, async Task, etc.) - Perfect compatibility with Unity workflows: supports hot-updated MonoBehaviour, ScriptableObject, DOTS technology, and correct instantiation of scripts attached to assets - High-performance register-based interpreter, with performance metrics significantly better than other solutions - Memory efficiency: hot-update classes occupy the same memory as regular C# classes - Original DHE (Differential Hybrid Execution) technology: unmodified functions run via AOT, changed functions run via interpreter, achieving near-native performance - Supports hot reload (100% assembly unloading), hotfix (without restart), and dll encryption - Supports MonoPInvokeCallback and interaction with languages like Lua/JS/Python It supports Unity 2019.4 through 6000.x full LTS series, as well as Tuanjie Engine and HarmonyOS platforms. Thousands of commercial projects have adopted it, with over a hundred iOS games in the top 500 free charts using it, covering genres such as MMORPG, heavy card games, and tower defense. It is widely used by leading domestic game companies. The project is licensed under MIT, providing complete documentation, quick-start guides, commercial cases, and commercial support. The author, walon, is the founder of Code Philosophy and graduated from Tsinghua University's Physics Department. Related projects include LeanCLR (a lightweight cross-platform CLR implementation), ZLua, ZenTS, and Obfuz.