paper-search -nodejs

Paper Search MCP (Node.js) is a Node.js-based MCP server that aggregates search, download, and full-text access capabilities across 14 academic platforms. It centralizes diverse sources such as arXiv, Web of Science, PubMed, Google Scholar, Sci-Hub, ScienceDirect, Springer, Wiley, Scopus, Crossref, and more, providing a unified data model and a consistent interface for performing searches and obtaining papers. You can run the MCP server via NPX (recommended) or locally with a built server file, and then connect to it through Claude Desktop or compatible MCP clients to issue search_papers, search_crossref, search_arxiv, and related actions. The server handles platform selection, rate limiting, and error handling to deliver relevant academic results while sanitizing inputs and validating DOIs when downloading content.

Key capabilities include:

• search_papers: perform cross-platform searches with options to target specific platforms or allow random selection for efficiency.
• search_crossref: perform Crossref-based searches by keyword, year, author, and sorting preferences (default to Crossref).
• search_arxiv: targeted searches for arXiv preprints.
• download_paper and full-text access through supported platforms, with platform-specific constraints (e.g., Wiley may restrict keyword searches).
• API-key driven integrations: Web of Science, PubMed, Semantic Scholar, Elsevier/ScienceDirect, Springer, Wiley, and others require API keys configured via environment variables.
• Unified data model and TypeScript support for safer integration and tooling across your MCP workflow.

Prerequisites:\n- Node.js >= 18.0.0 (and npm or yarn)\n- Internet access to install npm packages and fetch MCP server dependencies\n\nStep-by-step installation:\n1) Clone the repository (replace with the actual repo URL):\nbash\ngit clone https://github.com/your-username/paper-search-mcp-nodejs.git\ncd paper-search-mcp-nodejs\n\n2) Install dependencies:\nbash\nnpm install\n\n3) Create and configure environment variables:\nbash\ncp .env.example .env\n Edit .env to include your API keys, e.g. WOS_API_KEY, PubMed key, Elsevier key, Springer keys, Wiley token, etc.\n4) Build and run the server (NPX method):\nbash\nnpx -y paper-search-mcp-nodejs\n Or, install locally and run:

npm run build
npm start

5. If you prefer the local installation configuration for Claude Desktop, use the provided examples to point to the built server file afterwards. You can also run in development mode if provided by the project (npm run dev).

Tips and notes:\n- Ensure Web of Science, PubMed, Semantic Scholar, Elsevier (ScienceDirect/Scopus), Springer, Wiley, and other APIs have valid keys set in the environment variables. Some services require separate keys or tokens and may impose rate limits.\n- Wiley support may be limited to DOI-based PDF downloads; keyword search may not be supported for Wiley via some endpoints.\n- If using NPX deployment, the server runs on-demand without a global install. For persistent usage or Claude Desktop integration, consider the Local Installation configuration and point your claude_desktop_config.json to the built server.js path.\n- When using cross-platform search, you can specify platform filters like crossref, arxiv, webofscience, pubmed, biorxiv, medrxiv, semantic, iacr, googlescholar, scihub, sciencedirect, springer, scopus, or wiley. Be mindful of platform-specific capabilities and API key requirements.\n- Keep .env secure and do not commit keys to version control. Consider using environment variable management or secret managers in production.