The ArXiv MCP Server is a tool designed to facilitate AI assistants in accessing and searching for research papers on the arXiv repository. By using the Message Control Protocol (MCP), it provides a seamless interface for querying papers with specific filters, downloading content, and managing local storage of papers for efficient access. This server is particularly useful for developers and researchers who want to integrate arXiv's vast repository of academic papers into their AI models or applications, enabling automated and programmatic access to scholarly content.
Aucun avis encore. Soyez le premier à donner votre avis !
Connectez-vous pour rejoindre la conversation
Highlight AISearch arXiv by query, category, date, and sort order. Returns up to 5 compact results by default.
Fetch metadata and an abstract by arXiv ID. Does not download the paper.
Download and convert a paper to local Markdown. HTML first; PDF fallback uses '[pdf]'; can force re-fetch.
List papers stored locally. Returns id, title, authors, published; use 'compact' for IDs only.
Read locally stored paper content. Bounded to 12,000 chars by default; supports 'start', 'max_chars', 'return_full_text'.
Retrieve a paginated markdown heading outline of a paper.
Read one bounded markdown section of a paper by outline ID or unique title.
Perform a bounded passage search in a paper. Returns source offsets.
Retrieve bounded author-submitted LaTeX from a paper.
Return a paginated LaTeX outline of a paper. Supports 'start' and 'max_sections'.
Read one bounded LaTeX section of a paper by outline ID or exact title.
Fetch references and citing papers using the Semantic Scholar API.
Export BibTeX for one or more arXiv IDs based on authoritative arXiv metadata.
Save or update an arXiv topic watch. Stored locally.
List saved topic watches. This is read-only and does not advance last_checked.
Check saved watches for new papers since the last check.
Delete a saved topic watch by exact topic match.
Search downloaded papers by semantic similarity. Requires '[pro]'.
Rebuild the local semantic index. Requires '[pro]'.
