MCP Svelte Docs Server

# mcp-svelte-docs A Model Context Protocol (MCP) server that provides efficient access to Svelte documentation with advanced caching, search capabilities, and optimised content delivery. This server integrates directly with Svelte's official documentation, offering both full and compressed variants suitable for different LLM context window sizes. <a href="https://glama.ai/mcp/servers/wu4hy1xtjb"> <img width="380" height="200" src="https://glama.ai/mcp/servers/wu4hy1xtjb/badge" /> </a> ## Features - 📚 Complete Svelte documentation access through MCP Resources - 🔍 Advanced search capabilities: - Document type filtering (API, Tutorial, Example, Error) - Section hierarchy awareness - Intelligent relevance scoring based on: - Term frequency - Section importance - Document type relevance - Exact phrase matching - Context-aware result excerpts - Related search suggestions - 💾 Efficient caching with LibSQL - 🔄 Automatic content freshness checks - 📦 Support for package-specific documentation (Svelte, Kit, CLI) - 📏 Smart content chunking for large documents - 🗜️ Compressed variants for smaller context windows - 🏗️ Built on the Model Context Protocol ## Configuration This server requires configuration through your MCP client. Here are examples for different environments: ### Cline Configuration Add this to your Cline MCP settings: ```json { "mcpServers": { "svelte-docs": { "command": "npx", "args": ["-y", "mcp-svelte-docs"], "env": { "LIBSQL_URL": "file:local.db", "LIBSQL_AUTH_TOKEN": "your-auth-token-if-using-remote-db" } } } } ``` ### Claude Desktop with WSL Configuration For WSL environments, add this to your Claude Desktop configuration: ```json { "mcpServers": { "svelte-docs": { "command": "wsl.exe", "args": [ "bash", "-c", "LIBSQL_URL=file:local.db LIBSQL_AUTH_TOKEN=your-token npx -y mcp-svelte-docs" ] } } } ``` ### Environment Variables The server supports the following environment variables: - `LIBSQL_URL` (optional): URL for the LibSQL database. Defaults to `file:local.db` - `LIBSQL_AUTH_TOKEN` (optional): Auth token for remote LibSQL database ## API The server implements both MCP Resources and Tools: ### Resources Access documentation through these URIs: - `svelte-docs://docs/llms.txt` - Documentation index - `svelte-docs://docs/llms-full.txt` - Complete documentation - `svelte-docs://docs/llms-small.txt` - Compressed documentation - `svelte-docs://docs/{package}/llms.txt` - Package-specific documentation - Supported packages: svelte, kit, cli ### Tools #### search_docs Enhanced search functionality with advanced filtering and context awareness. Parameters: - `query` (string, required): Search keywords or natural language query - `doc_type` (string, optional): Filter by documentation type - Values: 'api', 'tutorial', 'example', 'error', 'all' - Default: 'all' - `context` (number, optional): Number of surrounding paragraphs (0-3) - Default: 1 - `include_hierarchy` (boolean, optional): Include section hierarchy - Default: true Example Usage: ```json // API Reference Search { "query": "bind:value directive", "doc_type": "api", "context": 1 } // Tutorial Search { "query": "routing sveltekit", "doc_type": "tutorial", "context": 2, "include_hierarchy": true } ``` #### get_next_chunk Retrieve subsequent chunks of large documents. Parameters: - `uri` (string, required): Document URI - `chunk_number` (number, required): Chunk number to retrieve (1-based) ## Development ### Setup 1. Clone the repository 2. Install dependencies: ```bash pnpm install ``` 3. Build the project: ```bash pnpm build ``` 4. Run in development mode: ```bash pnpm dev ``` ### Publishing 1. Update version in package.json 2. Build the project: ```bash pnpm build ``` 3. Publish to npm: ```bash pnpm publish ``` ## Contributing Contributions are welcome! Please feel free to submit a Pull Request. ## License MIT License - see the [LICENSE](LICENSE) file for details. ## Acknowledgments - Built on the [Model Context Protocol](https://github.com/modelcontextprotocol) - Powered by [Svelte Documentation](https://svelte.dev) - Uses [LibSQL](https://github.com/libsql/libsql) for efficient caching