MCP PyBoy Emulator Server
A Model Context Protocol server that allows LLMs to interact with Game Boy games through PyBoy emulation, providing capabilities to load ROMs, control games, capture screens, save/load states, and maintain game knowledge.
README
MCP PyBoy Emulator Server
An MCP (Model Context Protocol) server that enables LLMs to interact with Game Boy games through PyBoy emulation.
Overview
This server provides LLMs with the ability to:
- 🎮 Load and play Game Boy ROM files
- 🎯 Control games through button inputs and sequences
- 📸 Capture and analyze game screens
- 💾 Save and load game states
- 📝 Maintain persistent knowledge about games
Prerequisites
- Python 3.10 or higher
- uv - Fast Python package manager
- A compatible LLM client that supports MCP protocol
Installation
1. Install uv (if not already installed)
# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
# Or using pip
pip install uv
2. Clone the repository
git clone https://github.com/yourusername/mcp-pyboy.git
cd mcp-pyboy
3. Set up the development environment
# Install all dependencies (creates .venv automatically)
uv sync
# Install with development dependencies
uv sync --extra dev
4. Verify installation
# Check the CLI works
uv run mcp-pyboy
# Run tests (when implemented)
uv run pytest
Quick Start
Running the Server
# Basic usage
uv run mcp-pyboy
# With options (coming soon)
uv run mcp-pyboy --roms-dir ./roms --log-level DEBUG
Project Structure
mcp-pyboy/
├── src/mcp_pyboy/ # Main package
│ ├── mcp_server/ # MCP protocol implementation
│ ├── game_session/ # PyBoy emulator wrapper
│ ├── notebook/ # Knowledge persistence
│ ├── handlers/ # MCP tool implementations
│ └── utils/ # Shared utilities
├── tests/ # Test suite
├── docs/ # Architecture documentation
├── roms/ # ROM files directory
├── saves/ # Save states directory
└── notebooks/ # Game knowledge storage
Development
Development Tools
This project uses modern Python development tools:
- uv - Fast dependency management
- Black - Code formatting (88 char line length)
- Ruff - Linting and import sorting
- MyPy - Static type checking
- pytest - Testing framework
Common Development Commands
# Format code
uv run black src/ tests/
# Run linter
uv run ruff check src/ tests/
# Type check
uv run mypy src/
# Run tests
uv run pytest tests/ -v
# Install new dependency
uv add <package-name>
# Install dev dependency
uv add --dev <package-name>
VS Code Integration
This project includes VS Code configuration for optimal development:
.vscode/settings.json- Workspace settings with Black/Ruff integration.vscode/extensions.json- Recommended extensions
The configuration ensures:
- Black handles all formatting
- Ruff handles linting only (no formatting conflicts)
- Proper Python interpreter from virtual environment
MCP Tools Available
Once fully implemented, the server will provide these tools:
Emulation Control
load_rom- Load a Game Boy ROM filereset_game- Reset the current gameset_emulation_speed- Control game speed
Input Control
press_button- Press a Game Boy buttonhold_button- Hold a button downrelease_button- Release a held buttonsend_input_sequence- Execute a sequence of inputs
Screen and State
capture_screen- Get current game screensave_state- Save current game stateload_state- Load a saved statelist_states- List available save states
Knowledge Management
create_note- Create a note about the gameupdate_note- Update existing notesearch_notes- Search game knowledgelist_notes- List all notes for current game
Architecture
For detailed architecture documentation, see:
Contributing
This project is under active development. See the MVP Roadmap for current progress and planned features.
License
[License information to be added]
Acknowledgments
- Built on PyBoy - Game Boy emulator
- Uses MCP - Model Context Protocol
- Developed with Claude Code assistance
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。