LocalLlama

LocalLlama is a cutting-edge Unity package that wraps OllamaSharp, enabling AI integration in Unity ECS projects. It's designed for developers looking to incorporate multi-agent systems for development assistance and runtime interactions, such as game mastering or NPC dialogues.

Features

• Local and Remote Execution: Run llama2 AI locally or via client-server architecture.
• Multi-Agent System: Support for multiple AI agents.
• Development Tools: Code authoring, project editing, testing, and troubleshooting within Unity.
• In-Game Console: Access AI functionalities at runtime through an in-game console.
• Editor Integration: Use menu commands and panels for easy access in the Unity editor.
• C# Binding: Includes a C# DLL for direct interaction with OllamaSharp.

Installation

1. Open your Unity Project
2. Navigate to the Package Manage
3. click on the + icon in top left
4. select install with git url
5. paste https://github.com/cat-game-research/LocalLlama.git into the field.

Development

1. Clone the repository:

   git clone https://github.com/yourusername/LocalLlama.git
   

2. Run the setup script:

   bin\setup.bat
   

Usage

Add LocalLlama as a package in Unity, and it will be downloaded from GitHub. Utilize the in-game terminal or editor menus to interact with the AI. For detailed usage, refer to the Examples directory.

Contributing

Contributions are welcome! Please adhere to the Unity Style Guide.

License

LocalLlama is available under the MIT License. See LICENSE.md for more details.

Support

For support or feedback, join our Discord community (coming soon).

Citation

If you use LocalLlama in your research or project, please cite as follows:

@misc{LocalLlama,
  author = {cat-game-research},
  title = {LocalLlama - Unity Package and API for local Llama2 models using OllamaSharp},
  year = {2024},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\\url{https://github.com/yourusername/LocalLlama}}
}