[{"_1":2,"_335":-5,"_336":-5},"loaderData",{"_3":4,"_29":30,"_31":32,"_275":276},"root",{"_5":6,"_14":15},"toasts",[7],{"_8":9,"_10":11,"_12":13},"uid","01972642-22cc-77e0-8b2d-5c237c15796e","title","This MCP server does not support inspection","type","error","visitor",{"_16":17,"_18":19},"viewportWidth",1280,"visitorSession",{"_20":21,"_23":24,"_25":-5,"_26":27,"_28":-5},"attributes",[22],"BOT","id",7058884,"membership","referenceId","01971406-43e3-74e8-b880-2f000258007a","userAccount","routes/_public/_layout",{},"routes/_public/mcp/servers/~namespace/~slug/_pages/_layout",{"_33":34,"_75":-5,"_76":35,"_36":77,"_193":194,"_274":149},"availableLocales",[35,47,54,61,68],{"_36":37,"_40":44,"_45":46},"mcpServer",{"_38":39,"_40":41,"_42":43},"description","A Model Context Protocol server that enables Claude to perform advanced web research with intelligent search queuing, enhanced content extraction, and deep research capabilities.","name","MCP Deep Web Research Server","readme","# MCP Deep Web Research Server (v0.3.0)\n\n[![Node.js Version](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org/)\n[![TypeScript](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n[![smithery badge](https://smithery.ai/badge/@PedroDnT/mcp-deepwebresearch)](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)\n\nA Model Context Protocol (MCP) server for advanced web research.\n\n## Latest Changes\n\n- Added visit_page tool for direct webpage content extraction\n- Optimized performance to work within MCP timeout limits\n * Reduced default maxDepth and maxBranching parameters\n * Improved page loading efficiency\n * Added timeout checks throughout the process\n * Enhanced error handling for timeouts\n\n> This project is a fork of [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) by [mzxrai](https://github.com/mzxrai), enhanced with additional features for deep web research capabilities. We're grateful to the original creators for their foundational work.\n\nBring real-time info into Claude with intelligent search queuing, enhanced content extraction, and deep research capabilities.\n\n## Features\n\n- Intelligent Search Queue System\n - Batch search operations with rate limiting\n - Queue management with progress tracking\n - Error recovery and automatic retries\n - Search result deduplication\n\n- Enhanced Content Extraction\n - TF-IDF based relevance scoring\n - Keyword proximity analysis\n - Content section weighting\n - Readability scoring\n - Improved HTML structure parsing\n - Structured data extraction\n - Better content cleaning and formatting\n\n- Core Features\n - Google search integration\n - Webpage content extraction\n - Research session tracking\n - Markdown conversion with improved formatting\n\n## Prerequisites\n\n- [Node.js](https://nodejs.org/) >= 18 (includes `npm` and `npx`)\n- [Claude Desktop app](https://claude.ai/download)\n\n## Installation\n\n### Installing via Smithery\n\nTo install Deep Web Research Server for Claude Desktop automatically via [Smithery](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch):\n\n```bash\nnpx -y @smithery/cli install @PedroDnT/mcp-deepwebresearch --client claude\n```\n\n### Global Installation (Recommended)\n\n```bash\n# Install globally using npm\nnpm install -g mcp-deepwebresearch\n\n# Or using yarn\nyarn global add mcp-deepwebresearch\n\n# Or using pnpm\npnpm add -g mcp-deepwebresearch\n```\n\n### Local Project Installation\n\n```bash\n# Using npm\nnpm install mcp-deepwebresearch\n\n# Using yarn\nyarn add mcp-deepwebresearch\n\n# Using pnpm\npnpm add mcp-deepwebresearch\n```\n\n### Claude Desktop Integration\n\nAfter installing the package, add this entry to your `claude_desktop_config.json`:\n\n#### Windows\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\nLocation: `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n#### macOS\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\nLocation: `~/Library/Application Support/Claude/claude_desktop_config.json`\n\nThis config allows Claude Desktop to automatically start the web research MCP server when needed.\n\n### First-time Setup\n\nAfter installation, run this command to install required browser dependencies:\n```bash\nnpx playwright install chromium\n```\n\n## Usage\n\nSimply start a chat with Claude and send a prompt that would benefit from web research. If you'd like a prebuilt prompt customized for deeper web research, you can use the `agentic-research` prompt that we provide through this package. Access that prompt in Claude Desktop by clicking the Paperclip icon in the chat input and then selecting `Choose an integration` → `deepwebresearch` → `agentic-research`.\n\n### Tools\n\n1. `deep_research`\n - Performs comprehensive research with content analysis\n - Arguments:\n ```typescript\n {\n topic: string;\n maxDepth?: number; // default: 2\n maxBranching?: number; // default: 3\n timeout?: number; // default: 55000 (55 seconds)\n minRelevanceScore?: number; // default: 0.7\n }\n ```\n - Returns:\n ```typescript\n {\n findings: {\n mainTopics: Array<{name: string, importance: number}>;\n keyInsights: Array<{text: string, confidence: number}>;\n sources: Array<{url: string, credibilityScore: number}>;\n };\n progress: {\n completedSteps: number;\n totalSteps: number;\n processedUrls: number;\n };\n timing: {\n started: string;\n completed?: string;\n duration?: number;\n operations?: {\n parallelSearch?: number;\n deduplication?: number;\n topResultsProcessing?: number;\n remainingResultsProcessing?: number;\n total?: number;\n };\n };\n }\n ```\n\n2. `parallel_search`\n - Performs multiple Google searches in parallel with intelligent queuing\n - Arguments: `{ queries: string[], maxParallel?: number }`\n - Note: maxParallel is limited to 5 to ensure reliable performance\n\n3. `visit_page`\n - Visit a webpage and extract its content\n - Arguments: `{ url: string }`\n - Returns:\n ```typescript\n {\n url: string;\n title: string;\n content: string; // Markdown formatted content\n }\n ```\n\n### Prompts\n\n#### `agentic-research`\nA guided research prompt that helps Claude conduct thorough web research. The prompt instructs Claude to:\n- Start with broad searches to understand the topic landscape\n- Prioritize high-quality, authoritative sources\n- Iteratively refine the research direction based on findings\n- Keep you informed and let you guide the research interactively\n- Always cite sources with URLs\n\n## Configuration Options\n\nThe server can be configured through environment variables:\n\n- `MAX_PARALLEL_SEARCHES`: Maximum number of concurrent searches (default: 5)\n- `SEARCH_DELAY_MS`: Delay between searches in milliseconds (default: 200)\n- `MAX_RETRIES`: Number of retry attempts for failed requests (default: 3)\n- `TIMEOUT_MS`: Request timeout in milliseconds (default: 55000)\n- `LOG_LEVEL`: Logging level (default: 'info')\n\n## Error Handling\n\n### Common Issues\n\n1. Rate Limiting\n - Symptom: \"Too many requests\" error\n - Solution: Increase `SEARCH_DELAY_MS` or decrease `MAX_PARALLEL_SEARCHES`\n\n2. Network Timeouts\n - Symptom: \"Request timed out\" error\n - Solution: Ensure requests complete within the 60-second MCP timeout\n\n3. Browser Issues\n - Symptom: \"Browser failed to launch\" error\n - Solution: Ensure Playwright is properly installed (`npx playwright install`)\n\n### Debugging\n\nThis is beta software. If you run into issues:\n\n1. Check Claude Desktop's MCP logs:\n ```bash\n # On macOS\n tail -n 20 -f ~/Library/Logs/Claude/mcp*.log\n \n # On Windows\n Get-Content -Path \"$env:APPDATA\\Claude\\logs\\mcp*.log\" -Tail 20 -Wait\n ```\n\n2. Enable debug logging:\n ```bash\n export LOG_LEVEL=debug\n ```\n\n## Development\n\n### Setup\n\n```bash\n# Install dependencies\npnpm install\n\n# Build the project\npnpm build\n\n# Watch for changes\npnpm watch\n\n# Run in development mode\npnpm dev\n```\n\n### Testing\n\n```bash\n# Run all tests\npnpm test\n\n# Run tests in watch mode\npnpm test:watch\n\n# Run tests with coverage\npnpm test:coverage\n```\n\n### Code Quality\n\n```bash\n# Run linter\npnpm lint\n\n# Fix linting issues\npnpm lint:fix\n\n# Type check\npnpm type-check\n```\n\n## Contributing\n\n1. Fork the repository\n2. Create your feature branch (`git checkout -b feature/amazing-feature`)\n3. Commit your changes (`git commit -m 'Add some amazing feature'`)\n4. Push to the branch (`git push origin feature/amazing-feature`)\n5. Open a Pull Request\n\n### Coding Standards\n\n- Follow TypeScript best practices\n- Maintain test coverage above 80%\n- Document new features and APIs\n- Update CHANGELOG.md for significant changes\n- Follow semantic versioning\n\n### Performance Considerations\n\n- Use batch operations where possible\n- Implement proper error handling and retries\n- Consider memory usage with large datasets\n- Cache results when appropriate\n- Use streaming for large content\n\n## Requirements\n\n- Node.js >= 18\n- Playwright (automatically installed as a dependency)\n\n## Verified Platforms\n\n- [x] macOS\n- [x] Windows\n- [ ] Linux\n\n## License\n\nMIT\n\n## Credits\n\nThis project builds upon the excellent work of [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) by [mzxrai](https://github.com/mzxrai). The original codebase provided the foundation for our enhanced features and capabilities.\n\n## Author\n\n[qpd-v](https://github.com/qpd-v)","English","tag","en-US",{"_36":48,"_40":52,"_45":53},{"_38":49,"_40":50,"_42":51},"Un servidor de protocolo de contexto modelo que permite a Claude realizar investigaciones web avanzadas con colas de búsqueda inteligentes, extracción de contenido mejorada y capacidades de investigación profunda.","Servidor de investigación web profunda MCP","# Servidor de investigación web profunda MCP (v0.3.0)\n\n[![Versión de Node.js](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org/)[![Mecanografiado](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/) [![Licencia: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [![insignia de herrería](https://smithery.ai/badge/@PedroDnT/mcp-deepwebresearch)](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)\n\nUn servidor de Protocolo de Contexto Modelo (MCP) para investigación web avanzada.\n\n## Últimos cambios\n\n* Se agregó la herramienta visit\\_page para la extracción directa de contenido de la página web\n* Rendimiento optimizado para trabajar dentro de los límites de tiempo de espera de MCP\n * Parámetros maxDepth y maxBranching predeterminados reducidos\n * Eficiencia de carga de páginas mejorada\n * Se agregaron controles de tiempo de espera durante todo el proceso.\n * Manejo mejorado de errores en tiempos de espera\n\n> Este proyecto es una bifurcación de [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) de [mzxrai](https://github.com/mzxrai) , mejorada con funciones adicionales para la investigación en la web profunda. Agradecemos a los creadores originales su labor fundacional.\n\nIncorpore información en tiempo real a Claude con colas de búsqueda inteligentes, extracción de contenido mejorada y capacidades de investigación profunda.\n\n## Características\n\n* Sistema de cola de búsqueda inteligente\n\n * Operaciones de búsqueda por lotes con limitación de velocidad\n * Gestión de colas con seguimiento del progreso\n * Recuperación de errores y reintentos automáticos\n * Desduplicación de resultados de búsqueda\n\n* Extracción de contenido mejorada\n\n * Puntuación de relevancia basada en TF-IDF\n * Análisis de proximidad de palabras clave\n * Ponderación de la sección de contenido\n * Puntuación de legibilidad\n * Análisis mejorado de la estructura HTML\n * Extracción de datos estructurados\n * Mejor limpieza y formato de contenido\n\n* Características principales\n\n * Integración de búsqueda de Google\n * Extracción de contenido de páginas web\n * Seguimiento de sesiones de investigación\n * Conversión de Markdown con formato mejorado\n\n## Prerrequisitos\n\n* [Node.js](https://nodejs.org/) >= 18 (incluye `npm` y `npx` )\n* [Aplicación de escritorio Claude](https://claude.ai/download)\n\n## Instalación\n\n### Instalación mediante herrería\n\nPara instalar Deep Web Research Server para Claude Desktop automáticamente a través de [Smithery](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch) :\n\n```bash\nnpx -y @smithery/cli install @PedroDnT/mcp-deepwebresearch --client claude\n```\n\n### Instalación global (recomendada)\n\n```bash\n# Install globally using npm\nnpm install -g mcp-deepwebresearch\n\n# Or using yarn\nyarn global add mcp-deepwebresearch\n\n# Or using pnpm\npnpm add -g mcp-deepwebresearch\n```\n\n### Instalación de proyecto local\n\n```bash\n# Using npm\nnpm install mcp-deepwebresearch\n\n# Using yarn\nyarn add mcp-deepwebresearch\n\n# Using pnpm\npnpm add mcp-deepwebresearch\n```\n\n### Integración de escritorio de Claude\n\nDespués de instalar el paquete, agregue esta entrada a su `claude_desktop_config.json` :\n\n#### Ventanas\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\nUbicación: `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n#### macOS\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\nUbicación: `~/Library/Application Support/Claude/claude_desktop_config.json`\n\nEsta configuración permite que Claude Desktop inicie automáticamente el servidor de investigación web MCP cuando sea necesario.\n\n### Configuración por primera vez\n\nDespués de la instalación, ejecute este comando para instalar las dependencias necesarias del navegador:\n\n```bash\nnpx playwright install chromium\n```\n\n## Uso\n\nSimplemente inicia un chat con Claude y envía una propuesta que se beneficie de una investigación web. Si deseas una propuesta prediseñada y personalizada para una investigación web más profunda, puedes usar la propuesta `agentic-research` que ofrecemos en este paquete. Accede a esa propuesta en Claude Desktop haciendo clic en el icono del clip en la entrada del chat y seleccionando \" `Choose an integration` → `deepwebresearch` → `agentic-research` .\n\n### Herramientas\n\n1. `deep_research`\n\n * Realiza una investigación exhaustiva con análisis de contenido.\n\n * Argumentos:\n\n ```typescript\n {\n topic: string;\n maxDepth?: number; // default: 2\n maxBranching?: number; // default: 3\n timeout?: number; // default: 55000 (55 seconds)\n minRelevanceScore?: number; // default: 0.7\n }\n ```\n\n * Devoluciones:\n\n ```typescript\n {\n findings: {\n mainTopics: Array<{name: string, importance: number}>;\n keyInsights: Array<{text: string, confidence: number}>;\n sources: Array<{url: string, credibilityScore: number}>;\n };\n progress: {\n completedSteps: number;\n totalSteps: number;\n processedUrls: number;\n };\n timing: {\n started: string;\n completed?: string;\n duration?: number;\n operations?: {\n parallelSearch?: number;\n deduplication?: number;\n topResultsProcessing?: number;\n remainingResultsProcessing?: number;\n total?: number;\n };\n };\n }\n ```\n\n2. `parallel_search`\n\n * Realiza múltiples búsquedas de Google en paralelo con cola inteligente\n * Argumentos: `{ queries: string[], maxParallel?: number }`\n * Nota: maxParallel está limitado a 5 para garantizar un rendimiento confiable\n\n3. `visit_page`\n\n * Visita una página web y extrae su contenido\n\n * Argumentos: `{ url: string }`\n\n * Devoluciones:\n\n ```typescript\n {\n url: string;\n title: string;\n content: string; // Markdown formatted content\n }\n ```\n\n### Indicaciones\n\n#### `agentic-research`\n\nUna guía de investigación que ayuda a Claude a realizar una investigación web exhaustiva. La guía le indica a Claude que:\n\n* Comience con búsquedas amplias para comprender el panorama temático.\n* Priorizar fuentes confiables y de alta calidad\n* Refinar iterativamente la dirección de la investigación en función de los hallazgos\n* Manténgase informado y permítanos guiar la investigación de forma interactiva.\n* Cite siempre las fuentes con URL\n\n## Opciones de configuración\n\nEl servidor se puede configurar a través de variables de entorno:\n\n* `MAX_PARALLEL_SEARCHES` : Número máximo de búsquedas simultáneas (predeterminado: 5)\n* `SEARCH_DELAY_MS` : Retraso entre búsquedas en milisegundos (valor predeterminado: 200)\n* `MAX_RETRIES` : Número de reintentos para solicitudes fallidas (valor predeterminado: 3)\n* `TIMEOUT_MS` : Tiempo de espera de la solicitud en milisegundos (valor predeterminado: 55000)\n* `LOG_LEVEL` : Nivel de registro (predeterminado: 'info')\n\n## Manejo de errores\n\n### Problemas comunes\n\n1. Limitación de velocidad\n\n * Síntoma: Error \"Demasiadas solicitudes\"\n * Solución: Aumente `SEARCH_DELAY_MS` o disminuya `MAX_PARALLEL_SEARCHES`\n\n2. Tiempos de espera de la red\n\n * Síntoma: Error \"Tiempo de espera agotado\"\n * Solución: Asegúrese de que las solicitudes se completen dentro del tiempo de espera de MCP de 60 segundos\n\n3. Problemas con el navegador\n\n * Síntoma: Error \"No se pudo iniciar el navegador\"\n * Solución: asegúrese de que Playwright esté instalado correctamente ( `npx playwright install` )\n\n### Depuración\n\nEste software es beta. Si tiene algún problema:\n\n1. Consulte los registros MCP de Claude Desktop:\n\n ```bash\n # On macOS\n tail -n 20 -f ~/Library/Logs/Claude/mcp*.log\n\n # On Windows\n Get-Content -Path \"$env:APPDATA\\Claude\\logs\\mcp*.log\" -Tail 20 -Wait\n ```\n\n2. Habilitar el registro de depuración:\n\n ```bash\n export LOG_LEVEL=debug\n ```\n\n## Desarrollo\n\n### Configuración\n\n```bash\n# Install dependencies\npnpm install\n\n# Build the project\npnpm build\n\n# Watch for changes\npnpm watch\n\n# Run in development mode\npnpm dev\n```\n\n### Pruebas\n\n```bash\n# Run all tests\npnpm test\n\n# Run tests in watch mode\npnpm test:watch\n\n# Run tests with coverage\npnpm test:coverage\n```\n\n### Calidad del código\n\n```bash\n# Run linter\npnpm lint\n\n# Fix linting issues\npnpm lint:fix\n\n# Type check\npnpm type-check\n```\n\n## Contribuyendo\n\n1. Bifurcar el repositorio\n2. Crea tu rama de funciones ( `git checkout -b feature/amazing-feature` )\n3. Confirme sus cambios ( `git commit -m 'Add some amazing feature'` )\n4. Empujar a la rama ( `git push origin feature/amazing-feature` )\n5. Abrir una solicitud de extracción\n\n### Estándares de codificación\n\n* Siga las mejores prácticas de TypeScript\n* Mantener la cobertura de pruebas por encima del 80%\n* Documentar nuevas funciones y API\n* Actualice CHANGELOG.md para cambios significativos\n* Seguir el versionado semántico\n\n### Consideraciones de rendimiento\n\n* Utilice operaciones por lotes siempre que sea posible\n* Implementar un manejo adecuado de errores y reintentos\n* Considere el uso de memoria con grandes conjuntos de datos\n* Almacenar en caché los resultados cuando sea apropiado\n* Utilice la transmisión para contenido de gran tamaño\n\n## Requisitos\n\n* Node.js >= 18\n* Dramaturgo (instalado automáticamente como dependencia)\n\n## Plataformas verificadas\n\n* \\[x] macOS\n* \\[x] Ventanas\n* \\[ ] Linux\n\n## Licencia\n\nInstituto Tecnológico de Massachusetts (MIT)\n\n## Créditos\n\nEste proyecto se basa en el excelente trabajo de [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) de [mzxrai](https://github.com/mzxrai) . El código base original sentó las bases para nuestras funciones y capacidades mejoradas.\n\n## Autor\n\n[qpd-v](https://github.com/qpd-v)","Spanish","es-ES",{"_36":55,"_40":59,"_45":60},{"_38":56,"_40":57,"_42":58},"Claude がインテリジェントな検索キューイング、強化されたコンテンツ抽出、および詳細な調査機能を使用して高度な Web 調査を実行できるようにするモデルコンテキストプロトコルサーバー。","MCP ディープウェブリサーチサーバー","# MCP ディープウェブリサーチサーバー (v0.3.0)\n\n[![Node.js バージョン](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org/) [![タイプスクリプト](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/) [![ライセンス: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [![鍛冶屋のバッジ](https://smithery.ai/badge/@PedroDnT/mcp-deepwebresearch)](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)\n\n高度な Web 調査のためのモデルコンテキストプロトコル (MCP) サーバー。\n\n## 最新の変更点\n\n* ウェブページコンテンツを直接抽出するための visit\\_page ツールを追加しました\n* MCPタイムアウト制限内で動作するように最適化されたパフォーマンス\n * デフォルトのmaxDepthとmaxBranchingパラメータを削減\n * ページの読み込み効率の向上\n * プロセス全体にタイムアウトチェックを追加しました\n * タイムアウト時のエラー処理の強化\n\n> このプロジェクトは、 [mzxrai](https://github.com/mzxrai)による[mcp-webresearch](https://github.com/mzxrai/mcp-webresearch)のフォークであり、ディープウェブリサーチ機能のための追加機能が強化されています。私たちは、元の作成者たちの基礎的な作業に感謝いたします。\n\nインテリジェントな検索キューイング、強化されたコンテンツ抽出、詳細な調査機能により、Claude にリアルタイムの情報を取り込みます。\n\n## 特徴\n\n* インテリジェント検索キューシステム\n\n * レート制限付きのバッチ検索操作\n * 進捗状況を追跡するキュー管理\n * エラー回復と自動再試行\n * 検索結果の重複排除\n\n* 強化されたコンテンツ抽出\n\n * TF-IDFベースの関連性スコアリング\n * キーワード近接分析\n * コンテンツセクションの重み付け\n * 読みやすさスコア\n * HTML構造解析の改善\n * 構造化データ抽出\n * コンテンツの整理とフォーマットの改善\n\n* コア機能\n\n * Google検索統合\n * ウェブページコンテンツの抽出\n * 研究セッションの追跡\n * フォーマットが改善されたマークダウン変換\n\n## 前提条件\n\n* [Node.js](https://nodejs.org/) >= 18 ( `npm`と`npx`を含む)\n* [クロードデスクトップアプリ](https://claude.ai/download)\n\n## インストール\n\n### Smithery経由でインストール\n\n[Smithery](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)経由で Claude Desktop 用の Deep Web Research Server を自動的にインストールするには:\n\n```bash\nnpx -y @smithery/cli install @PedroDnT/mcp-deepwebresearch --client claude\n```\n\n### グローバルインストール（推奨）\n\n```bash\n# Install globally using npm\nnpm install -g mcp-deepwebresearch\n\n# Or using yarn\nyarn global add mcp-deepwebresearch\n\n# Or using pnpm\npnpm add -g mcp-deepwebresearch\n```\n\n### ローカルプロジェクトのインストール\n\n```bash\n# Using npm\nnpm install mcp-deepwebresearch\n\n# Using yarn\nyarn add mcp-deepwebresearch\n\n# Using pnpm\npnpm add mcp-deepwebresearch\n```\n\n### クロードデスクトップ統合\n\nパッケージをインストールした後、 `claude_desktop_config.json`に次のエントリを追加します。\n\n#### ウィンドウズ\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n場所: `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n#### macOS\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n場所: `~/Library/Application Support/Claude/claude_desktop_config.json`\n\nこの設定により、Claude Desktop は必要に応じて Web リサーチ MCP サーバーを自動的に起動できるようになります。\n\n### 初回セットアップ\n\nインストール後、このコマンドを実行して必要なブラウザ依存関係をインストールします。\n\n```bash\nnpx playwright install chromium\n```\n\n## 使用法\n\nClaude とチャットを開始し、Web リサーチに役立つプロンプトを送信するだけです。より詳細な Web リサーチ向けにカスタマイズされた既成のプロンプトが必要な場合は、このパッケージで提供されている`agentic-research`プロンプトをご利用ください。Claude Desktop でこのプロンプトにアクセスするには、チャット入力欄のペーパークリップアイコンをクリックし、 `Choose an integration` → `deepwebresearch` → `agentic-research`を選択します。\n\n### ツール\n\n1. `deep_research`\n\n * コンテンツ分析による包括的な調査を実施\n\n * 引数:\n\n ```typescript\n {\n topic: string;\n maxDepth?: number; // default: 2\n maxBranching?: number; // default: 3\n timeout?: number; // default: 55000 (55 seconds)\n minRelevanceScore?: number; // default: 0.7\n }\n ```\n\n * 戻り値:\n\n ```typescript\n {\n findings: {\n mainTopics: Array<{name: string, importance: number}>;\n keyInsights: Array<{text: string, confidence: number}>;\n sources: Array<{url: string, credibilityScore: number}>;\n };\n progress: {\n completedSteps: number;\n totalSteps: number;\n processedUrls: number;\n };\n timing: {\n started: string;\n completed?: string;\n duration?: number;\n operations?: {\n parallelSearch?: number;\n deduplication?: number;\n topResultsProcessing?: number;\n remainingResultsProcessing?: number;\n total?: number;\n };\n };\n }\n ```\n\n2. `parallel_search`\n\n * インテリジェントなキューイングで複数の Google 検索を並行して実行します\n * 引数: `{ queries: string[], maxParallel?: number }`\n * 注: 信頼性の高いパフォーマンスを確保するために、maxParallel は 5 に制限されています。\n\n3. `visit_page`\n\n * ウェブページにアクセスしてコンテンツを抽出する\n\n * 引数: `{ url: string }`\n\n * 戻り値:\n\n ```typescript\n {\n url: string;\n title: string;\n content: string; // Markdown formatted content\n }\n ```\n\n### プロンプト\n\n#### `agentic-research`\n\nクロードが徹底的なウェブ調査を行うのに役立つガイド付きリサーチプロンプト。このプロンプトは、クロードに以下の指示を与えます。\n\n* トピックの状況を理解するために、まずは広範囲な検索から始めましょう\n* 高品質で信頼できる情報源を優先する\n* 調査結果に基づいて研究の方向性を繰り返し改善する\n* 情報を提供し、インタラクティブに研究を進めることができます\n* 常にURLでソースを引用する\n\n## 設定オプション\n\nサーバーは環境変数を通じて設定できます:\n\n* `MAX_PARALLEL_SEARCHES` : 同時検索の最大数（デフォルト: 5）\n* `SEARCH_DELAY_MS` : 検索間の遅延（ミリ秒単位）（デフォルト: 200）\n* `MAX_RETRIES` : 失敗したリクエストの再試行回数（デフォルト: 3）\n* `TIMEOUT_MS` : リクエストタイムアウト（ミリ秒）（デフォルト: 55000）\n* `LOG_LEVEL` : ログレベル（デフォルト: 'info'）\n\n## エラー処理\n\n### よくある問題\n\n1. レート制限\n\n * 症状: 「リクエストが多すぎます」というエラー\n * 解決策: `SEARCH_DELAY_MS`を増やすか、 `MAX_PARALLEL_SEARCHES`を減らす\n\n2. ネットワークタイムアウト\n\n * 症状: 「リクエストがタイムアウトしました」というエラー\n * 解決策: リクエストが60秒のMCPタイムアウト内に完了することを確認する\n\n3. ブラウザの問題\n\n * 症状: 「ブラウザの起動に失敗しました」というエラー\n * 解決策: Playwright が正しくインストールされていることを確認してください ( `npx playwright install` )\n\n### デバッグ\n\nこれはベータ版ソフトウェアです。問題が発生した場合は、以下の手順に従ってください。\n\n1. Claude Desktop の MCP ログを確認します。\n\n ```bash\n # On macOS\n tail -n 20 -f ~/Library/Logs/Claude/mcp*.log\n\n # On Windows\n Get-Content -Path \"$env:APPDATA\\Claude\\logs\\mcp*.log\" -Tail 20 -Wait\n ```\n\n2. デバッグログを有効にする:\n\n ```bash\n export LOG_LEVEL=debug\n ```\n\n## 発達\n\n### 設定\n\n```bash\n# Install dependencies\npnpm install\n\n# Build the project\npnpm build\n\n# Watch for changes\npnpm watch\n\n# Run in development mode\npnpm dev\n```\n\n### テスト\n\n```bash\n# Run all tests\npnpm test\n\n# Run tests in watch mode\npnpm test:watch\n\n# Run tests with coverage\npnpm test:coverage\n```\n\n### コード品質\n\n```bash\n# Run linter\npnpm lint\n\n# Fix linting issues\npnpm lint:fix\n\n# Type check\npnpm type-check\n```\n\n## 貢献\n\n1. リポジトリをフォークする\n2. 機能ブランチを作成します（ `git checkout -b feature/amazing-feature` ）\n3. 変更をコミットします ( `git commit -m 'Add some amazing feature'` )\n4. ブランチにプッシュする ( `git push origin feature/amazing-feature` )\n5. プルリクエストを開く\n\n### コーディング標準\n\n* TypeScriptのベストプラクティスに従う\n* テストカバレッジを80%以上維持する\n* 新しい機能とAPIを文書化する\n* 重要な変更についてはCHANGELOG.mdを更新してください\n* セマンティックバージョニングに従う\n\n### パフォーマンスに関する考慮事項\n\n* 可能な場合はバッチ操作を使用する\n* 適切なエラー処理と再試行を実装する\n* 大規模なデータセットでのメモリ使用量を考慮する\n* 適切な場合に結果をキャッシュする\n* 大容量コンテンツにはストリーミングを使用する\n\n## 要件\n\n* Node.js >= 18\n* Playwright (依存関係として自動的にインストールされます)\n\n## 検証済みプラットフォーム\n\n* \\[x] macOS\n* \\[x] ウィンドウズ\n* \\[ ] リナックス\n\n## ライセンス\n\nマサチューセッツ工科大学\n\n## クレジット\n\nこのプロジェクトは、 [mzxrai](https://github.com/mzxrai)による[mcp-webresearch](https://github.com/mzxrai/mcp-webresearch)の優れた成果を基盤としています。オリジナルのコードベースは、私たちの強化された機能と性能の基盤となりました。\n\n## 著者\n\n[qpd-v](https://github.com/qpd-v)","Japanese","ja-JP",{"_36":62,"_40":66,"_45":67},{"_38":63,"_40":64,"_42":65},"클로드가 지능형 검색 대기열, 향상된 콘텐츠 추출, 심층적인 조사 기능을 통해 고급 웹 조사를 수행할 수 있도록 하는 모델 컨텍스트 프로토콜 서버입니다.","MCP 딥웹 리서치 서버","# MCP 딥웹 연구 서버(v0.3.0)\n\n[![Node.js 버전](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org/) [![타입스크립트](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/) [![라이센스: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [![대장간 배지](https://smithery.ai/badge/@PedroDnT/mcp-deepwebresearch)](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)\n\n고급 웹 연구를 위한 MCP(Model Context Protocol) 서버입니다.\n\n## 최신 변경 사항\n\n* 웹페이지 콘텐츠를 직접 추출하기 위한 visit\\_page 도구가 추가되었습니다.\n* MCP 시간 초과 제한 내에서 작동하도록 성능이 최적화되었습니다.\n * 기본 maxDepth 및 maxBranching 매개변수가 축소되었습니다.\n * 향상된 페이지 로딩 효율성\n * 프로세스 전반에 걸쳐 시간 초과 확인이 추가되었습니다.\n * 시간 초과에 대한 향상된 오류 처리\n\n> 이 프로젝트는 [mzxrai](https://github.com/mzxrai) 가 개발한 [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) 의 포크(fork)로, 딥웹 리서치 기능을 위한 추가 기능이 강화되었습니다. 초기 개발에 기여해 주신 개발자분들께 감사드립니다.\n\n지능형 검색 대기열, 향상된 콘텐츠 추출, 심층 조사 기능을 통해 클로드에 실시간 정보를 제공합니다.\n\n## 특징\n\n* 지능형 검색 대기열 시스템\n\n * 속도 제한을 사용한 일괄 검색 작업\n * 진행 상황 추적을 통한 대기열 관리\n * 오류 복구 및 자동 재시도\n * 검색 결과 중복 제거\n\n* 향상된 콘텐츠 추출\n\n * TF-IDF 기반 관련성 점수\n * 키워드 근접 분석\n * 콘텐츠 섹션 가중치\n * 가독성 점수\n * 개선된 HTML 구조 파싱\n * 구조화된 데이터 추출\n * 더 나은 콘텐츠 정리 및 서식\n\n* 핵심 기능\n\n * Google 검색 통합\n * 웹페이지 콘텐츠 추출\n * 연구 세션 추적\n * 향상된 서식을 사용한 마크다운 변환\n\n## 필수 조건\n\n* [Node.js](https://nodejs.org/) >= 18 ( `npm` 및 `npx` 포함)\n* [클로드 데스크톱 앱](https://claude.ai/download)\n\n## 설치\n\n### Smithery를 통해 설치\n\n[Smithery](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch) 를 통해 Claude Desktop용 Deep Web Research Server를 자동으로 설치하려면:\n\n지엑스피1\n\n### 글로벌 설치(권장)\n\n```bash\n# Install globally using npm\nnpm install -g mcp-deepwebresearch\n\n# Or using yarn\nyarn global add mcp-deepwebresearch\n\n# Or using pnpm\npnpm add -g mcp-deepwebresearch\n```\n\n### 로컬 프로젝트 설치\n\n```bash\n# Using npm\nnpm install mcp-deepwebresearch\n\n# Using yarn\nyarn add mcp-deepwebresearch\n\n# Using pnpm\npnpm add mcp-deepwebresearch\n```\n\n### Claude 데스크톱 통합\n\n패키지를 설치한 후 `claude_desktop_config.json` 에 다음 항목을 추가하세요.\n\n#### 윈도우\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n위치: `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n#### 맥OS\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n위치: `~/Library/Application Support/Claude/claude_desktop_config.json`\n\n이 구성을 사용하면 Claude Desktop이 필요할 때 웹 연구 MCP 서버를 자동으로 시작할 수 있습니다.\n\n### 첫 번째 설정\n\n설치 후 다음 명령을 실행하여 필요한 브라우저 종속성을 설치합니다.\n\n```bash\nnpx playwright install chromium\n```\n\n## 용법\n\nClaude와 채팅을 시작하고 웹 리서치에 도움이 될 만한 메시지를 보내세요. 심층적인 웹 리서치에 맞춰 미리 만들어진 프롬프트를 원하시면 이 패키지를 통해 제공되는 `agentic-research` 프롬프트를 사용하실 수 있습니다. Claude 데스크톱에서 채팅 입력란의 페이퍼클립 아이콘을 클릭한 다음, `Choose an integration` → `deepwebresearch` → `agentic-research` 선택하여 해당 프롬프트에 접속하세요.\n\n### 도구\n\n1. `deep_research`\n\n * 콘텐츠 분석을 통해 포괄적인 연구를 수행합니다.\n\n * 인수:\n\n ```typescript\n {\n topic: string;\n maxDepth?: number; // default: 2\n maxBranching?: number; // default: 3\n timeout?: number; // default: 55000 (55 seconds)\n minRelevanceScore?: number; // default: 0.7\n }\n ```\n\n * 보고:\n\n ```typescript\n {\n findings: {\n mainTopics: Array<{name: string, importance: number}>;\n keyInsights: Array<{text: string, confidence: number}>;\n sources: Array<{url: string, credibilityScore: number}>;\n };\n progress: {\n completedSteps: number;\n totalSteps: number;\n processedUrls: number;\n };\n timing: {\n started: string;\n completed?: string;\n duration?: number;\n operations?: {\n parallelSearch?: number;\n deduplication?: number;\n topResultsProcessing?: number;\n remainingResultsProcessing?: number;\n total?: number;\n };\n };\n }\n ```\n\n2. `parallel_search`\n\n * 지능형 대기열을 통해 여러 Google 검색을 병렬로 수행합니다.\n * 인수: `{ queries: string[], maxParallel?: number }`\n * 참고: maxParallel은 안정적인 성능을 보장하기 위해 5로 제한됩니다.\n\n3. `visit_page`\n\n * 웹 페이지를 방문하여 콘텐츠 추출\n\n * 인수: `{ url: string }`\n\n * 보고:\n\n ```typescript\n {\n url: string;\n title: string;\n content: string; // Markdown formatted content\n }\n ```\n\n### 프롬프트\n\n#### `agentic-research`\n\n클로드가 철저한 웹 조사를 수행하는 데 도움이 되는 가이드 조사 프롬프트입니다. 이 프롬프트는 클로드에게 다음을 지시합니다.\n\n* 주제 환경을 이해하려면 광범위한 검색으로 시작하세요.\n* 고품질의 권위 있는 출처를 우선시하세요\n* 연구 결과를 바탕으로 연구 방향을 반복적으로 개선합니다.\n* 귀하에게 정보를 제공하고 대화형으로 연구를 안내합니다.\n* 항상 URL을 사용하여 출처를 인용하세요\n\n## 구성 옵션\n\n서버는 환경 변수를 통해 구성할 수 있습니다.\n\n* `MAX_PARALLEL_SEARCHES` : 동시 검색 최대 수 (기본값: 5)\n* `SEARCH_DELAY_MS` : 검색 간 지연 시간(밀리초) (기본값: 200)\n* `MAX_RETRIES` : 실패한 요청에 대한 재시도 횟수(기본값: 3)\n* `TIMEOUT_MS` : 요청 시간 초과(밀리초) (기본값: 55000)\n* `LOG_LEVEL` : 로깅 레벨(기본값: 'info')\n\n## 오류 처리\n\n### 일반적인 문제\n\n1. 속도 제한\n\n * 증상: \"요청이 너무 많습니다\" 오류\n * 해결 방법: `SEARCH_DELAY_MS` 늘리거나 `MAX_PARALLEL_SEARCHES` 줄이세요.\n\n2. 네트워크 시간 초과\n\n * 증상: \"요청 시간 초과\" 오류\n * 솔루션: 요청이 60초 MCP 시간 초과 내에 완료되도록 보장\n\n3. 브라우저 문제\n\n * 증상: \"브라우저를 시작하지 못했습니다\" 오류\n * 해결 방법: Playwright가 제대로 설치되었는지 확인하세요( `npx playwright install` )\n\n### 디버깅\n\n베타 소프트웨어입니다. 문제가 발생할 경우:\n\n1. Claude Desktop의 MCP 로그를 확인하세요.\n\n ```bash\n # On macOS\n tail -n 20 -f ~/Library/Logs/Claude/mcp*.log\n\n # On Windows\n Get-Content -Path \"$env:APPDATA\\Claude\\logs\\mcp*.log\" -Tail 20 -Wait\n ```\n\n2. 디버그 로깅 활성화:\n\n ```bash\n export LOG_LEVEL=debug\n ```\n\n## 개발\n\n### 설정\n\n```bash\n# Install dependencies\npnpm install\n\n# Build the project\npnpm build\n\n# Watch for changes\npnpm watch\n\n# Run in development mode\npnpm dev\n```\n\n### 테스트\n\n```bash\n# Run all tests\npnpm test\n\n# Run tests in watch mode\npnpm test:watch\n\n# Run tests with coverage\npnpm test:coverage\n```\n\n### 코드 품질\n\n```bash\n# Run linter\npnpm lint\n\n# Fix linting issues\npnpm lint:fix\n\n# Type check\npnpm type-check\n```\n\n## 기여하다\n\n1. 저장소를 포크하세요\n2. 기능 브랜치를 생성합니다( `git checkout -b feature/amazing-feature` )\n3. 변경 사항을 커밋하세요( `git commit -m 'Add some amazing feature'` )\n4. 브랜치에 푸시( `git push origin feature/amazing-feature` )\n5. 풀 리퀘스트 열기\n\n### 코딩 표준\n\n* TypeScript 모범 사례를 따르세요\n* 테스트 커버리지를 80% 이상으로 유지\n* 새로운 기능 및 API 문서화\n* 중요한 변경 사항이 있는 경우 CHANGELOG.md를 업데이트하세요.\n* 의미적 버전 관리를 따르세요\n\n### 성능 고려 사항\n\n* 가능한 경우 일괄 작업을 사용하세요\n* 적절한 오류 처리 및 재시도를 구현합니다.\n* 대용량 데이터 세트의 메모리 사용량을 고려하세요\n* 적절한 경우 캐시 결과\n* 대용량 콘텐츠에 스트리밍을 사용하세요\n\n## 요구 사항\n\n* 노드.js >= 18\n* Playwright(종속성으로 자동 설치됨)\n\n## 검증된 플랫폼\n\n* \\[x] 맥OS\n* \\[x] 윈도우\n* \\[ ] 리눅스\n\n## 특허\n\nMIT\n\n## 크레딧\n\n이 프로젝트는 [mzxrai](https://github.com/mzxrai) 가 개발한 [mcp-webresearch](https://github.com/mzxrai/mcp-webresearch) 의 훌륭한 작업을 기반으로 합니다. 원본 코드베이스는 향상된 기능과 성능의 기반을 제공했습니다.\n\n## 작가\n\n[qpd-v](https://github.com/qpd-v)","Korean","ko-KR",{"_36":69,"_40":73,"_45":74},{"_38":70,"_40":71,"_42":72},"模型上下文协议服务器使 Claude 能够通过智能搜索排队、增强内容提取和深度研究功能执行高级网络研究。","MCP 深层网络研究服务器","# MCP 深层网络研究服务器 (v0.3.0)\n\n[![Node.js 版本](https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg)](https://nodejs.org/) [![TypeScript](https://img.shields.io/badge/TypeScript-5.0-blue.svg)](https://www.typescriptlang.org/) [![许可证：MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [![铁匠徽章](https://smithery.ai/badge/@PedroDnT/mcp-deepwebresearch)](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)\n\n用于高级网络研究的模型上下文协议 (MCP) 服务器。\n\n## 最新变更\n\n* 增加visit\\_page工具，可直接提取网页内容\n* 优化性能以在 MCP 超时限制内工作\n * 减少默认的 maxDepth 和 maxBranching 参数\n * 提高页面加载效率\n * 在整个过程中添加了超时检查\n * 增强超时错误处理\n\n> 该项目是[mzxrai](https://github.com/mzxrai)开发的[mcp-webresearch](https://github.com/mzxrai/mcp-webresearch)的一个分支，并添加了深度网络研究的附加功能。我们感谢原作者的奠基性工作。\n\n通过智能搜索排队、增强内容提取和深度研究功能将实时信息带入 Claude。\n\n## 特征\n\n* 智能搜索队列系统\n\n * 具有速率限制的批量搜索操作\n * 带进度跟踪的队列管理\n * 错误恢复和自动重试\n * 搜索结果去重\n\n* 增强内容提取\n\n * 基于 TF-IDF 的相关性评分\n * 关键词接近度分析\n * 内容部分权重\n * 可读性评分\n * 改进的 HTML 结构解析\n * 结构化数据提取\n * 更好的内容清理和格式化\n\n* 核心功能\n\n * Google 搜索集成\n * 网页内容提取\n * 研究会话跟踪\n * 改进格式的 Markdown 转换\n\n## 先决条件\n\n* [Node.js](https://nodejs.org/) >= 18（包括`npm`和`npx` ）\n* [克劳德桌面应用程序](https://claude.ai/download)\n\n## 安装\n\n### 通过 Smithery 安装\n\n要通过[Smithery](https://smithery.ai/server/@PedroDnT/mcp-deepwebresearch)自动为 Claude Desktop 安装 Deep Web Research Server：\n\n```bash\nnpx -y @smithery/cli install @PedroDnT/mcp-deepwebresearch --client claude\n```\n\n### 全局安装（推荐）\n\n```bash\n# Install globally using npm\nnpm install -g mcp-deepwebresearch\n\n# Or using yarn\nyarn global add mcp-deepwebresearch\n\n# Or using pnpm\npnpm add -g mcp-deepwebresearch\n```\n\n### 本地项目安装\n\n```bash\n# Using npm\nnpm install mcp-deepwebresearch\n\n# Using yarn\nyarn add mcp-deepwebresearch\n\n# Using pnpm\npnpm add mcp-deepwebresearch\n```\n\n### Claude 桌面集成\n\n安装软件包后，将此条目添加到您的`claude_desktop_config.json`中：\n\n#### 视窗\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n位置： `%APPDATA%\\Claude\\claude_desktop_config.json`\n\n#### macOS\n\n```json\n{\n \"mcpServers\": {\n \"deepwebresearch\": {\n \"command\": \"mcp-deepwebresearch\",\n \"args\": []\n }\n }\n}\n```\n\n位置： `~/Library/Application Support/Claude/claude_desktop_config.json`\n\n此配置允许 Claude Desktop 在需要时自动启动网络研究 MCP 服务器。\n\n### 首次设置\n\n安装后，运行此命令安装所需的浏览器依赖项：\n\n```bash\nnpx playwright install chromium\n```\n\n## 用法\n\n只需与 Claude 开始聊天，并发送一个有助于进行网络研究的提示即可。如果您想要一个为更深入的网络研究而定制的预建提示，您可以使用我们通过此包提供的`agentic-research`提示。在 Claude Desktop 中，点击聊天输入框中的回形针图标，然后选择`Choose an integration` → `deepwebresearch` → `agentic-research`访问该提示。\n\n### 工具\n\n1. `deep_research`\n\n * 通过内容分析进行全面研究\n\n * 参数：\n\n ```typescript\n {\n topic: string;\n maxDepth?: number; // default: 2\n maxBranching?: number; // default: 3\n timeout?: number; // default: 55000 (55 seconds)\n minRelevanceScore?: number; // default: 0.7\n }\n ```\n\n * 返回：\n\n ```typescript\n {\n findings: {\n mainTopics: Array<{name: string, importance: number}>;\n keyInsights: Array<{text: string, confidence: number}>;\n sources: Array<{url: string, credibilityScore: number}>;\n };\n progress: {\n completedSteps: number;\n totalSteps: number;\n processedUrls: number;\n };\n timing: {\n started: string;\n completed?: string;\n duration?: number;\n operations?: {\n parallelSearch?: number;\n deduplication?: number;\n topResultsProcessing?: number;\n remainingResultsProcessing?: number;\n total?: number;\n };\n };\n }\n ```\n\n2. `parallel_search`\n\n * 通过智能排队并行执行多个 Google 搜索\n * 参数： `{ queries: string[], maxParallel?: number }`\n * 注意：maxParallel 限制为 5，以确保可靠的性能\n\n3. `visit_page`\n\n * 访问网页并提取其内容\n\n * 参数： `{ url: string }`\n\n * 返回：\n\n ```typescript\n {\n url: string;\n title: string;\n content: string; // Markdown formatted content\n }\n ```\n\n### 提示\n\n#### `agentic-research`\n\n引导式研究提示，帮助 Claude 进行深入的网络研究。该提示指导 Claude 执行以下操作：\n\n* 从广泛的搜索开始，了解主题概况\n* 优先考虑高质量、权威的来源\n* 根据研究结果反复完善研究方向\n* 让您随时了解情况，并让您以交互方式指导研究\n* 始终引用带有 URL 的来源\n\n## 配置选项\n\n可以通过环境变量配置服务器：\n\n* `MAX_PARALLEL_SEARCHES` ：最大并发搜索数（默认值：5）\n* `SEARCH_DELAY_MS` ：搜索之间的延迟时间（以毫秒为单位）（默认值：200）\n* `MAX_RETRIES` ：失败请求的重试次数（默认值：3）\n* `TIMEOUT_MS` ：请求超时（以毫秒为单位）（默认值：55000）\n* `LOG_LEVEL` ：日志记录级别（默认值：'info'）\n\n## 错误处理\n\n### 常见问题\n\n1. 速率限制\n\n * 症状：“请求过多”错误\n * 解决方案：增加`SEARCH_DELAY_MS`或减少`MAX_PARALLEL_SEARCHES`\n\n2. 网络超时\n\n * 症状：“请求超时”错误\n * 解决方案：确保请求在 60 秒 MCP 超时内完成\n\n3. 浏览器问题\n\n * 症状：“浏览器启动失败”错误\n * 解决方案：确保 Playwright 已正确安装（ `npx playwright install` ）\n\n### 调试\n\n这是测试版软件。如果您遇到问题：\n\n1. 检查 Claude Desktop 的 MCP 日志：\n\n ```bash\n # On macOS\n tail -n 20 -f ~/Library/Logs/Claude/mcp*.log\n\n # On Windows\n Get-Content -Path \"$env:APPDATA\\Claude\\logs\\mcp*.log\" -Tail 20 -Wait\n ```\n\n2. 启用调试日志记录：\n\n ```bash\n export LOG_LEVEL=debug\n ```\n\n## 发展\n\n### 设置\n\n```bash\n# Install dependencies\npnpm install\n\n# Build the project\npnpm build\n\n# Watch for changes\npnpm watch\n\n# Run in development mode\npnpm dev\n```\n\n### 测试\n\n```bash\n# Run all tests\npnpm test\n\n# Run tests in watch mode\npnpm test:watch\n\n# Run tests with coverage\npnpm test:coverage\n```\n\n### 代码质量\n\n```bash\n# Run linter\npnpm lint\n\n# Fix linting issues\npnpm lint:fix\n\n# Type check\npnpm type-check\n```\n\n## 贡献\n\n1. 分叉存储库\n2. 创建你的功能分支（ `git checkout -b feature/amazing-feature` ）\n3. 提交您的更改（ `git commit -m 'Add some amazing feature'` ）\n4. 推送到分支（ `git push origin feature/amazing-feature` ）\n5. 打开拉取请求\n\n### 编码标准\n\n* 遵循 TypeScript 最佳实践\n* 保持测试覆盖率在 80% 以上\n* 记录新功能和 API\n* 更新 CHANGELOG.md 以了解重大变化\n* 遵循语义版本控制\n\n### 性能考虑\n\n* 尽可能使用批处理操作\n* 实施适当的错误处理和重试\n* 考虑大型数据集的内存使用情况\n* 在适当的时候缓存结果\n* 使用流式传输来获取大量内容\n\n## 要求\n\n* Node.js >= 18\n* Playwright（作为依赖项自动安装）\n\n## 已验证的平台\n\n* \\[x] macOS\n* \\[x] Windows\n* Linux\n\n## 执照\n\n麻省理工学院\n\n## 致谢\n\n该项目以[mzxrai](https://github.com/mzxrai)的[mcp-webresearch](https://github.com/mzxrai/mcp-webresearch)的出色工作为基础。原始代码库为我们增强的功能和性能奠定了基础。\n\n## 作者\n\n[qpd-v](https://github.com/qpd-v)","Chinese","zh-CN","githubUser","locale",{"_78":79,"_80":81,"_20":109,"_118":39,"_119":39,"_120":121,"_124":125,"_146":-5,"_23":147,"_148":149,"_150":151,"_40":41,"_169":138,"_170":171,"_175":-5,"_176":-5,"_177":-5,"_178":179,"_184":-5,"_158":134,"_185":186,"_189":-5,"_8":190,"_191":192},"addedAt","2025-03-21T11:19:12.046046644Z","argumentsJsonSchema",{"_82":83,"_106":107,"_12":108},"properties",{"_84":85,"_90":91,"_94":95,"_98":99,"_102":103},"LOG_LEVEL",{"_86":87,"_38":88,"_12":89},"default","info","Logging level","string","TIMEOUT_MS",{"_86":92,"_38":93,"_12":89},"55000","Request timeout in milliseconds","MAX_RETRIES",{"_86":96,"_38":97,"_12":89},"3","Number of retry attempts for failed requests","SEARCH_DELAY_MS",{"_86":100,"_38":101,"_12":89},"200","Delay between searches in milliseconds","MAX_PARALLEL_SEARCHES",{"_86":104,"_38":105,"_12":89},"5","Maximum number of concurrent searches","required",[],"object",[110,111,112,113,114,115,116,117],"category:web-scraping","category:agent-orchestration","category:search","environment:network","hosting:hybrid","language:typescript","license_grade:a","author:claimed","descriptionMarkdown","descriptionPlainText","githubProject",{"_42":43,"_122":123},"url","https://github.com/pedrodnt/mcp-deepwebresearch","githubRepository",{"_126":127,"_128":129,"_130":131,"_132":133,"_40":134,"_135":136,"_140":141,"_142":143,"_145":141},"defaultBranch","main","fullName","PedroDnT/mcp-DEEPwebresearch","githubTreeId",97352,"language","TypeScript","mcp-DEEPwebresearch","owner",{"_137":138,"_8":139},"login","PedroDnT","VQXVcpvKdI","recentStargazers",0,"spdxLicense",{"_40":144},"MIT License","stargazers","glamaJson",2376,"inspectable",false,"integrations",[152,162],{"_153":154,"_38":160,"_8":161},"brand",{"_155":156,"_40":157,"_158":159},"logoSvgPath","M12.48 10.92v3.28h7.84c-.24 1.84-.853 3.187-1.787 4.133-1.147 1.147-2.933 2.4-6.053 2.4-4.827 0-8.6-3.893-8.6-8.72s3.773-8.72 8.6-8.72c2.6 0 4.507 1.027 5.907 2.347l2.307-2.307C18.747 1.44 16.133 0 12.48 0 5.867 0 .307 5.387.307 12s5.56 12 12.173 12c3.573 0 6.267-1.173 8.373-3.36 2.16-2.16 2.84-5.213 2.84-7.667 0-.76-.053-1.467-.173-2.053H12.48z","Google","slug","google","Integrates with Google Search to perform searches, with features for parallel searching, batch operations with rate limiting, and intelligent queuing of search requests.","ucvb7woi1i",{"_153":163,"_38":167,"_8":168},{"_155":164,"_40":165,"_158":166},"M22.27 19.385H1.73A1.73 1.73 0 010 17.655V6.345a1.73 1.73 0 011.73-1.73h20.54A1.73 1.73 0 0124 6.345v11.308a1.73 1.73 0 01-1.73 1.731zM5.769 15.923v-4.5l2.308 2.885 2.307-2.885v4.5h2.308V8.078h-2.308l-2.307 2.885-2.308-2.885H3.46v7.847zM21.232 12h-2.309V8.077h-2.307V12h-2.308l3.461 4.039z","Markdown","markdown","Converts web content to Markdown format with improved formatting, making research results more readable within Claude.","44o6a1q5u2","namespace","npmPackage",{"_40":172,"_173":174},"mcp-deepwebresearch","weeklyDownloads",53,"prompts","redditPostForeignId","resources","scores",{"_180":181,"_182":-5,"_183":-5},"license",100,"quality","security","simpleIcon","supportedPlatforms",[187,188],"MACOS","WINDOWS","tools","vsfb013k8i","updatedAt","2025-03-21T18:22:22.202202830Z","meta",[195,197,199,206,209,211,214,217,220,223,226,228,231,234,236,238,240,242,244],{"_10":196},"MCP Deep Web Research Server | Glama",{"_198":39,"_40":38},"content",{"_200":201,"_202":203,"_204":205},"href","https://glama.ai/mcp/servers/@PedroDnT/mcp-DEEPwebresearch","rel","canonical","tagName","link",{"_198":41,"_207":208},"property","og:title",{"_198":39,"_207":210},"og:description",{"_198":212,"_207":213},"https://glama.ai/logo.png","og:logo",{"_198":215,"_207":216},"https://glama.ai/mcp/servers/@PedroDnT/mcp-DEEPwebresearch/og-image","og:image",{"_198":218,"_207":219},"630","og:image:height",{"_198":221,"_207":222},"1200","og:image:width",{"_198":224,"_207":225},"website","og:type",{"_198":201,"_207":227},"og:url",{"_198":229,"_207":230},"Glama – MCP Hosting Platform","og:site_name",{"_198":232,"_40":233},"summary_large_image","twitter:card",{"_198":229,"_40":235},"twitter:title",{"_198":39,"_40":237},"twitter:description",{"_198":215,"_40":239},"twitter:image",{"_198":218,"_40":241},"twitter:image:height",{"_198":221,"_40":243},"twitter:image:width",{"_245":246},"script:ld+json",{"_247":248,"_249":250},"@context","http://schema.org","@graph",[251],{"_247":252,"_253":254,"_255":256},"https://schema.org","@type","BreadcrumbList","itemListElement",[257,264,268,272],{"_253":258,"_259":260,"_40":261,"_262":263},"ListItem","item","https://glama.ai/","Glama","position",1,{"_253":258,"_259":265,"_40":266,"_262":267},"https://glama.ai/mcp","MCP",2,{"_253":258,"_259":269,"_40":270,"_262":271},"https://glama.ai/mcp/servers","Servers",3,{"_253":258,"_259":201,"_40":41,"_262":273},4,"visitorIsMaintainer","routes/_public/mcp/servers/~namespace/~slug/_pages/reviews/_route",{"_75":-5,"_36":277,"_193":299,"_334":-5},{"_78":79,"_80":278,"_20":286,"_118":39,"_119":39,"_120":287,"_124":288,"_146":-5,"_23":147,"_148":149,"_150":291,"_40":41,"_169":138,"_170":296,"_175":-5,"_176":-5,"_177":-5,"_178":297,"_184":-5,"_158":134,"_185":298,"_189":-5,"_8":190,"_191":192},{"_82":279,"_106":285,"_12":108},{"_84":280,"_90":281,"_94":282,"_98":283,"_102":284},{"_86":87,"_38":88,"_12":89},{"_86":92,"_38":93,"_12":89},{"_86":96,"_38":97,"_12":89},{"_86":100,"_38":101,"_12":89},{"_86":104,"_38":105,"_12":89},[],[110,111,112,113,114,115,116,117],{"_42":43,"_122":123},{"_126":127,"_128":129,"_130":131,"_132":133,"_40":134,"_135":289,"_140":141,"_142":290,"_145":141},{"_137":138,"_8":139},{"_40":144},[292,294],{"_153":293,"_38":160,"_8":161},{"_155":156,"_40":157,"_158":159},{"_153":295,"_38":167,"_8":168},{"_155":164,"_40":165,"_158":166},{"_40":172,"_173":174},{"_180":181,"_182":-5,"_183":-5},[187,188],[300,302,304,306,308,309,310,311,312,313,314,315,316,317,318,319,320,321,322],{"_10":301},"Reviews | MCP Deep Web Research Server | Glama",{"_198":303,"_40":38},"Read community reviews and feedback for MCP Deep Web Research Server. Discover user experiences, ratings, strengths, and potential limitations of this MCP implementation from actual users.",{"_200":305,"_202":203,"_204":205},"https://glama.ai/mcp/servers/@PedroDnT/mcp-DEEPwebresearch/reviews",{"_198":307,"_207":208},"Reviews | MCP Deep Web Research Server",{"_198":303,"_207":210},{"_198":212,"_207":213},{"_198":215,"_207":216},{"_198":218,"_207":219},{"_198":221,"_207":222},{"_198":224,"_207":225},{"_198":305,"_207":227},{"_198":229,"_207":230},{"_198":232,"_40":233},{"_198":229,"_40":235},{"_198":303,"_40":237},{"_198":215,"_40":239},{"_198":218,"_40":241},{"_198":221,"_40":243},{"_245":323},{"_247":248,"_249":324},[325],{"_247":252,"_253":254,"_255":326},[327,328,329,330,331],{"_253":258,"_259":260,"_40":261,"_262":263},{"_253":258,"_259":265,"_40":266,"_262":267},{"_253":258,"_259":269,"_40":270,"_262":271},{"_253":258,"_259":201,"_40":41,"_262":273},{"_253":258,"_259":305,"_40":332,"_262":333},"Reviews",5,"review","actionData","errors"]

Rate & Review

MCP directory API