Chess-com MCP (interactive views)
Chess.com player, game, and daily-puzzle tools where each tool ships its own interactive React view — board replays and a playable puzzle widget, not just text. Built with Skybridge for ChatGPT & Claude.
README
<p align="center"> <img src="./public/demo-banner.png" alt="chess-mcp demo" /> </p>
<p align="center"> <a href="https://docs.skybridge.tech"><img src="https://img.shields.io/badge/Skybridge-1.0.4-2563eb?style=flat-square&logo=react&logoColor=white" alt="Skybridge" /></a> <a href="https://chess.niklas.sh"><img src="https://img.shields.io/badge/Demo-chess.niklas.sh-7c3aed?style=flat-square&logo=lichess&logoColor=white" alt="Live demo" /></a> <a href="./LICENSE"><img src="https://img.shields.io/badge/license-Beerware-fbbf24?style=flat-square" alt="License: Beerware" /></a> </p>
A reference example for building MCP apps with interactive React views — where each tool ships with its own view instead of returning plain text. Built with Skybridge, it looks up Chess.com players, games, and the daily puzzle, and runs in ChatGPT and Claude.
Try it
- Live demo: chess.niklas.sh
- MCP server URL:
https://mcp.chess.niklas.sh/mcp
Just add the MCP server URL to any compatible client (ChatGPT, Claude, etc.) — that's it. Then ask things like "show me magnuscarlsen's last game" or "give me today's chess puzzle".
<p align="center"> <img src="./public/demo-get_daily_puzzle_claude.gif" alt="Solving the daily puzzle in Claude" width="600" /> </p>
Tools
| Tool | Input | Description |
|---|---|---|
| get-chess-player | username |
Player profile with rapid, blitz, and bullet ratings plus win/loss/draw records. |
| get-last-game | username |
The player's most recent game — result, opponent, opening, and an interactive board replay. |
| get-daily-puzzle | none | The Chess.com daily puzzle as an interactive board: solve it move by move with live feedback, reveal the solution, or reset and retry. |
Getting Started
Prerequisites
- Node.js 24.14.1+
- pnpm 9+
Install
pnpm install
Start the dev server
pnpm dev
This starts:
- The MCP server at
http://localhost:3000/mcp. - The Skybridge DevTools UI at
http://localhost:3000.
All scripts (dev, build, test, lint, typecheck) are defined in
package.json.
Project Structure
├── src/
│ ├── server.ts # Server entry: tool definitions + analytics middleware
│ ├── chess-com.ts # Typed Chess.com API client (zod boundary + helpers)
│ ├── analytics.ts # PostHog wrapper (no-op without a key)
│ ├── helpers.ts # Typed useToolInfo / useCallTool hooks
│ ├── views/ # One React view per tool
│ │ └── shared/ # Shared view code (chess board, helpers)
│ └── index.css # Global styles
├── vite.config.ts # Vite + Skybridge + Tailwind config
├── Dockerfile # Cloud Run image
└── package.json
Testing
Test the app locally using the DevTools UI at http://localhost:3000 while running pnpm dev.
Unit tests run with Vitest: pnpm test.
To connect with web clients like ChatGPT or Claude, expose your server with the --tunnel flag (pnpm dev:tunnel). See the test guide.
Deployment & Analytics
This app deploys to Google Cloud Run via release-please, and tool calls can be
tracked with PostHog (no-op unless configured). See
docs/deployment.md for the full setup.
Resources
- Skybridge Documentation
- Apps SDK Documentation
- MCP Apps Documentation
- Model Context Protocol Documentation
License
The source code is released under the Beerware License.
The chess piece icons are from the "Chess" pack on Flaticon and are used under the Flaticon Free License — they are not covered by the Beerware license. See NOTICE for details.
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。