component-mcp-server
Exposes design system component documentation to LLM clients via MCP, allowing listing and retrieving component contexts from Markdown files with YAML frontmatter.
README
component-mcp-server
A standalone Model Context Protocol (MCP) server that exposes our design system's component documentation to LLM clients.
Component docs live as Markdown files (with YAML frontmatter) in contexts/. The server reads them at request time — no build step is needed to add or update a component.
Tools
list_components
No input. Scans the contexts/ directory for all .md files and returns their names (without the .md extension) as a JSON array. Use this to discover what components are documented before requesting one.
get_component_context
Input: { componentName: string }
Looks up contexts/<componentName>.md (case-insensitive). If found, returns:
frontmatter— the parsed YAML frontmatter as structured JSONbody— the raw markdown body (frontmatter stripped)raw— the full original file content (frontmatter + body together), so a caller can just read it as plain documentation
If the component isn't found, returns an error object listing the currently available component names so the caller can retry with a valid one.
Running locally
npm install
npm run dev
npm run dev runs the server directly via tsx (fast iteration, no build step). The server communicates over stdio, so it won't print anything on its own — it's meant to be driven by an MCP client.
Other scripts:
npm run build # compile TypeScript to dist/
npm run start # run the compiled server (node dist/index.js)
Testing standalone with MCP Inspector
npx @modelcontextprotocol/inspector npx tsx src/index.ts
This opens a web UI where you can call list_components and get_component_context directly and inspect the responses.
Connecting to Claude Code
claude mcp add --transport stdio component-mcp -- npx tsx <absolute path to src/index.ts>
Replace <absolute path to src/index.ts> with the full path on your machine, e.g. C:\Users\you\path\to\component-mcp-server\src\index.ts. Once added, Claude Code can call list_components and get_component_context directly.
Adding a new component
Drop a new <name>.md file into contexts/, following the same frontmatter + markdown body structure as contexts/button.md. No code changes are required — list_components and get_component_context will pick it up automatically.
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。