docs: Add comprehensive CLAUDE.md for development guidance

- Add essential development commands and environment setup instructions
- Document monorepo architecture with client/server structure
- Explain MCP server data processing pipeline and caching strategy
- Include internationalization and testing guidelines
- Provide performance considerations and deployment context

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

Author: Lei Xu

CLAUDE.md

@@ -0,0 +1,119 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Common Development Commands
+
+### Environment Setup
+- **Node Version Management**: This project requires Node.js 22.16.0 (specified in `.nvmrc`)
+  - Always run `nvm use` when switching to this project
+  - Use `./setup.sh` for automated environment setup including nvm and dependencies
+  - Alternative: manually run `nvm use` followed by `npm ci`
+
+### Development Commands
+- `npm run dev` - Start both client and server in development mode (uses concurrently)
+- `npm run client:dev` - Start only the React client (Vite dev server)
+- `npm run server:dev` - Start only the Express server with tsx watch mode
+- `npm run build` - Build both client and server for production
+- `npm run lint` - Run ESLint across the entire project
+- `npm run preview` - Preview production build of client
+
+### Testing
+- `cd server && npm run test` - Run all tests with Vitest
+- `cd server && npm run test:watch` - Run tests in watch mode
+- `cd server && npm run test:integration` - Run integration tests only
+- `cd server && npm run test:mock` - Run mock/unit tests only
+
+### Data Processing Scripts (Server)
+- `cd server && npm run crawl-servers` - Crawl MCP servers from the main repository
+- `cd server && npm run crawl-servers-postprocess` - Post-process crawled server data
+- `cd server && npm run clean-duplicates` - Remove duplicate entries
+- `cd server && npm run process_categories` - Process and organize category data
+- `cd server && npm run process_locales` - Process internationalization data
+- `cd server && npm run process_githubinfo` - Enrich servers with GitHub metadata
+
+## Architecture Overview
+
+### Project Structure
+This is a **monorepo workspace** with two main packages:
+- **client/**: React frontend (Vite + TypeScript + Tailwind CSS)
+- **server/**: Express.js backend (TypeScript + Node.js)
+
+### Backend Architecture (server/)
+
+#### Core Server Files
+- **server.ts**: Main Express application with route registration
+- **routes/**: API route handlers
+  - `mcp.ts`: MCP server-related endpoints
+  - `hub.ts`: Hub-specific endpoints
+- **lib/**: Core business logic
+  - `mcpServers.ts`: MCP server data management with locale-aware caching
+  - `config.ts`: Environment configuration
+  - `githubEnrichment.ts`: GitHub API integration for server metadata
+  - `llm.ts` & `llmTools.ts`: LLM integration for data processing
+
+#### Data Management
+- **src/data/split/**: Individual JSON files for each MCP server (enables efficient caching)
+- **Caching Strategy**: In-memory cache with TTL (1 hour) and locale-specific storage
+- **Data Processing Pipeline**: Crawling → GitHub enrichment → Localization → Split storage
+
+#### Key Data Structures
+- **McpServer Interface**: Primary server data model with GitHub metadata
+- **Locale Support**: Multi-language support (en, de, es, ja, zh-hans, zh-hant)
+- **Server Types**: Reference Server, Official Integration, Community Server, Framework, Resource
+
+### Frontend Architecture (client/)
+
+#### Core Structure
+- **React Router**: Client-side routing with pages in `src/pages/`
+- **Context API**: `LanguageContext` for internationalization
+- **Component Structure**: Reusable components in `src/components/`
+- **Data Management**: Static imports from `src/data/servers.ts`
+
+#### Key Pages
+- **Home**: Main landing with server discovery
+- **Listing**: Categorized server browsing with filtering
+- **ServerDetails**: Detailed server information (uses hubId routing)
+- **Submit**: Server submission form
+
+### Data Flow
+
+#### Server Data Pipeline
+1. **Crawling**: `mcp_servers_crawler.ts` extracts servers from official MCP repository
+2. **GitHub Enrichment**: `process_githubinfo.ts` adds stars, forks, commits, license info
+3. **Localization**: `process_locales.ts` generates translated versions
+4. **Categorization**: `process_categories.ts` organizes servers by category
+5. **Split Storage**: Individual JSON files for efficient loading
+
+#### API Layer
+- **Locale-aware endpoints**: All server endpoints support locale parameter
+- **Caching**: Server-side caching with automatic TTL refresh
+- **Error Handling**: Graceful fallbacks to default locale
+
+## Development Guidelines
+
+### Working with MCP Server Data
+- Server data is split into individual files under `server/src/data/split/`
+- Always use the caching functions in `mcpServers.ts` rather than direct file access
+- When adding new server data, run the processing pipeline to maintain consistency
+- GitHub enrichment requires valid GitHub API access
+
+### Internationalization
+- Locale files are in `client/src/locale/` and category translations in `category-*.json`
+- Server data supports locale-specific descriptions and content
+- Always test with multiple locales when modifying server data
+
+### Testing Strategy
+- Integration tests verify GitHub API functionality and data processing
+- Mock tests for isolated unit testing
+- Use `vitest` for all server-side testing
+
+### Environment Variables
+- Server requires GitHub API token for enrichment processes
+- Configuration managed through `config.ts` with environment variable support
+- Production deployment uses Docker with external environment files
+
+### Performance Considerations
+- MCP server data uses in-memory caching to avoid file system overhead
+- Individual server files enable partial loading strategies
+- Client uses static imports for core server data to leverage Vite bundling