What does Pico8 Mcp do?
It lets LLM agents play real PICO-8 cartridges in deterministic lockstep, providing low-level control over the game state.
@smartfrog
基本情報
MCP server that lets LLM agents play real PICO-8 carts in deterministic lockstep: press buttons, step frames, inspect screenshots, read Lua globals, peek memory, reset/replay, and watch live in a browser.
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"pico8": {
"command": "node",
"args": [
"/path/to/pico8-mcp/dist/index.js"
],
"env": {
"PICO8_BIN": "/path/to/pico8"
}
}
}
}ツール一覧はまだありません
この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。
ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。
Pico8 Mcp is an MCP server that enables LLM agents to play real PICO-8 cartridges in deterministic lockstep. It allows agents to press buttons, step frames, inspect screenshots, read Lua globals, peek memory, reset or replay, and watch the action live in a browser.
Run the server and connect an MCP-compatible LLM client. The server exposes tools that let the agent press buttons, advance frames, capture screenshots, read and write Lua globals, inspect raw memory, reset the cart, replay sequences, and serve a live browser view.
It lets LLM agents play real PICO-8 cartridges in deterministic lockstep, providing low-level control over the game state.
It supports pressing buttons, stepping frames, inspecting screenshots, reading Lua globals, peeking memory, resetting/replaying, and streaming live video to a browser.
It is designed for LLM agents that need to interact with PICO-8 games programmatically.

Model Context Protocol Servers

Connect your AI assistant to your own YOUB endurance training account. Read recorded activities with full analysis, planned sessions, race goals, wellness (HRV, sleep, resting HR) and your athletic profile. Create or mov
不止于 CSS
🚀 The fast, Pythonic way to build MCP servers and clients.
Hindsight: Agent Memory That Learns
コメント