MCPlayerOne
Enables playing AI-generated text-based maze adventure games with synthwave themes. Create custom game worlds from any theme or URL, explore locations, collect items, and compete on leaderboards through natural language commands.
README

🚀 MCPlayerOne
Turn your imagination into an AI generated game
Welcome to the MCPlayerOne — a synthwave, maze-crawling, world-building adventure platform! Create, explore, and play in wild, AI-generated worlds. Powered by Next.js, Langflow, and your imagination.
<p align="center"><i>Explore your world: Example gameplay</i></p>
🎮 Connect & Play via MCP Tools
You can play instantly by connecting to the public MCP Game Server using your favorite agent tools—no server setup required!
🧪 Warp Preview (Terminal)
- Download and install Warp Preview
- Open Warp Preview
- Go to
Settings→Manage MCP servers→+Add - Paste the following into the "SSE endpoint URL" field:
https://mcplayerone.onrender.com/api/v1/mcp/sse - Save and connect!
- You can now interact with the MCP Game Server directly from Warp Preview.
🖥️ Cursor (AI IDE)
- Install Cursor IDE
- Add the MCP Game Server as a tool server:
- Open Cursor and go to Extensions/Integrations.
- Add a new MCP tool server with this config:
{ "MCPlayerOne": { "transportType": "sse", "url": "https://mcplayerone.onrender.com/api/v1/mcp/sse", "openapi": "https://mcplayerone.onrender.com/api/v1/mcp/openapi.json" } }
- Start playing! Use Cursor's chat or code tools to send MCP commands, automate moves, or analyze game data.
🤖 Claude (AI Chatbot/Desktop)
- Install Python and pip
- Install uv:
pip install uv - Install mcp-proxy:
uvx install mcp-proxy - Add this to your Claude Desktop config:
{ "mcpServers": { "MCPlayerOne": { "command": "uvx", "args": ["mcp-proxy", "https://mcplayerone.onrender.com/api/v1/mcp/sse"] } } } - Restart Claude Desktop and select the MCPlayerOne server to start playing!
🛠️ Troubleshooting
- 404 or connection errors? Make sure you're using the correct Render URL above.
- Agent not responding? Check that you're using the correct SSE endpoint and OpenAPI URL.
- Session timeouts? SSE sessions last for 1 hour of inactivity—just reconnect if needed.
🚦 Getting Started
Whether you want to jump right into a game or set up your own server, getting started with MCPlayerOne is easy:
To Play Instantly
- Use any of the methods above (Warp Preview, Cursor, Claude, etc.) to connect to the public MCP Game Server.
Once connected, tell the agent you would like to play an MCPlayerOne game.
<div align="center" style="display: flex; gap: 24px; justify-content: center;"> <div> <img src="public/images/claude-example-game-start.png" alt="Start a Game in Claude" width="320" height="320"/> <div align="center"><i>Start an MCPlayerOne Game</i></div> </div> <div> <img src="public/images/claude-example-choose-game.png" alt="Play a Game in Claude" width="320" height="320"/> <div align="center"><i>Pick a Game and Play</i></div> </div> </div>
Now, just tell the agent which game you'd like to play and optionally give yourself a username (the agent will choose one for you if you don't).
Create your own game
You can also create and generate your own game to play.
Have fun with it, point it at a website, maybe a GitHub repo, or just give it a theme. The more creative and imaginative you are the more interesting your game. This is where AI generation really shines.
For example, tell the agent something like "Let's create a new MCPlayerOne game, use github https://github.com/SonicDMG/mcp-game-server for the theme" or "Create a new MCPlayerOne game with a theme based on DOOM playing as a space marine".
Game creation takes about 2-3 minutes so please be patient.
🎲 Gameplay
🗝️ Available MCP Tools
These are the main MCP tools you (or your agent) can use to play. Most agents will choose actions based on your input or goals.
| Tool Name | What It Does |
|---|---|
| listStories | List all available game stories that can be played. |
| createGame | Generate a new game story, locations, items, and cover image by theme. |
| startGame | Start a new game session for a user in a specific story, or resume if started. |
| getGameState | Get the current state of the game (player and location) for a user and story. |
| lookAround | Get details about the player's current location (description, items, exits). |
| movePlayer | Move the player to a new location (by target ID) if valid. |
| takeItem | Pick up a specified item from the current location. |
| examineTarget | Get a detailed description of a specific item or feature in the current room. |
| getLeaderboard | Retrieve the leaderboard data for a specific story. |
| getStoryById | Get the details of a specific story by its logical ID. |
✨ Features
- 🪐 AI-Generated Worlds: Instantly create new stories and mazes
- 👾 ASCII & Pixel Art: Retro visuals and grid-based layouts
- 🌌 Leaderboard: Compete for glory, see winners and the fallen
- 🚀 Artifacts & Rooms: Collect, explore, and interact
- 🐉 Synthwave Theme: Neon colors, pixel dragons, and more
- 🧑💻 Powered by Langflow: Next-level AI integration
🛠️ Local Development
Want to run your own server or contribute? Follow these steps:
npm install
npm run dev
Open http://localhost:3000 to play locally.
- Edit
.envto set up API keys, endpoints, or tool access as needed. - For Astra DB, Langflow, or other integrations, follow the comments in
.env.example.
MCP Tool Setup (for local dev)
- For local dev, use:
http://localhost:3000/api/v1/mcp/sse(SSE endpoint)http://localhost:3000/api/v1/mcp/openapi.json(OpenAPI spec)
- All MCP tools are enabled by default in local dev.
🌟 Contributing
Pull requests are welcome! Open an issue, fork, and help us build the wildest worlds in the metaverse.
- Star the repo ⭐
🔗 Links
📑 Error & Status Code Reference (Addendum)
This section documents the main error and status codes used by the MCP Game Server API and SSE/JSON-RPC interface, for quick reference:
HTTP Status Codes
| Code | When Issued | Meaning |
|---|---|---|
| 200 | Successful request (may include errors in body for game logic) | OK |
| 400 | Malformed request, missing/invalid parameters | Bad Request |
| 403 | Action not allowed (e.g., unmet requirements) | Forbidden |
| 404 | Resource not found (player, story, location, item) | Not Found |
| 409 | Conflict (e.g., story with same ID exists) | Conflict |
| 500 | Internal server error | Server Error |
JSON-RPC Error Codes (SSE/Tool Proxy)
| Code | When Issued | Meaning |
|---|---|---|
| -32000 | Generic server error (proxy failure, non-JSON response, etc.) | Server error (JSON-RPC 2.0 reserved) |
| -32601 | Tool or method not found | Method not found (JSON-RPC 2.0) |
| -32602 | Invalid parameters for a tool/method | Invalid params (JSON-RPC 2.0) |
| -32603 | Internal JSON-RPC error | Internal error (JSON-RPC 2.0) |
Game Logic/Response Fields
| Field | When Present | Meaning |
|---|---|---|
| success | All responses | true if operation succeeded, false otherwise |
| error | On failure | Error message for the user/agent |
| hint | Sometimes | Optional hint for the user/agent |
| needsPlayer | Player not found | Indicates the client should call /start |
Notes
- Game logic errors (e.g., wrong answer, missing item) may return HTTP 200 with
success: falseand anerrormessage in the body. - JSON-RPC error codes are used in SSE and tool proxying for agent compatibility.
- See the OpenAPI spec for full schema details.
Thanks for visiting, explorer! May your mazes be twisty and your artifacts shiny. 🟪🟦🟩🟧🟨🟫
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。