ArXiv Academic Search & Discovery MCP

smithery/alexgenovese/arxiv-mcp

# arXiv Academic Search & Discovery MCP Server Connect your AI assistants and coding agents directly to the massive **arXiv API** database of scientific papers, preprints, and academic research. This high-performance, strictly-typed Model Context Protocol (MCP) server enables AI agents (such as Cursor, Claude Desktop, and VS Code) to search, retrieve, and read scientific articles seamlessly. ## ✨ Key Features - 🔍 **Advanced Chronological Search**: Search millions of articles by query and sort them chronologically (`submittedDate` or `lastUpdatedDate`) or by relevance. - 👤 **Full Author & Collaboration Extraction**: Retrieves complete lists of authors, including comments and citations. - 📄 **Direct Agent Access (HTML & PDF)**: Generates canonical links for both raw `.pdf` files and the newly supported **arXiv HTML view** (`https://arxiv.org/html/...`), which is significantly easier for LLM agents to scrape and read. - 📥 **Atom Category Feeds**: Monitor specific scientific domains (e.g. `cs.AI`, `quant-ph`) for the latest releases. - 🚀 **Lightweight & High-Speed**: Built on native ESM with high-performance XML parsing for near-instant responses. ## 🛠️ Available Tools - `arxiv_search`: Search arXiv papers with custom queries, pagination, and sorting filters. - `arxiv_get_paper`: Retrieve full, rich metadata for any paper by its arXiv ID (includes comments, DOI, and journal references). - `arxiv_get_pdf_url`: Fast access to canonical PDF and HTML URLs for a given paper. - `arxiv_get_feed`: Access RSS/Atom feeds by categories (e.g., computer science, physics). - `arxiv_query_authors`: Search papers published by a specific author. - `arxiv_related_papers`: Find relevant topics and keyword suggestions for a paper. ## 💻 Installation To run this stdio-based server locally with Smithery, simply run: ```bash npx smithery install alexgenovese/arxiv-mcp ```