Best MCP servers for academic research
Academic and literature research has a shape that general web search does not handle well: you need primary papers, full text rather than snippets, citation graphs to follow ideas backward and forward, and authoritative reference articles to ground definitions. An agent wired to the right servers can search a preprint archive, download and read a paper as clean markdown, trace who cited whom, and pull a Wikipedia or wiki article to anchor background, all without leaving the conversation. The right server depends on the source you trust: an open preprint archive for cutting-edge work, an encyclopedia for grounding, or a general wiki engine for domain-specific knowledge bases. The servers below cover those shapes, each a real MCP server with a verified, current install config.
arXiv
Joseph Blazick
A popular MCP server that lets agents search arXiv, download papers, and read their full text as markdown for research workflows.
The arXiv server searches papers, downloads them, reads full text as markdown, and supports semantic search and citation graphs, the core tool for working with preprints and tracing literature.
Wikipedia
Ravi Kumar E
A maintained MCP server that grounds an agent in Wikipedia: search articles, pull full text or summaries, sections, links, and key facts.
The Wikipedia server grounds an agent in articles, summaries, sections, links, and key facts, the fastest way to anchor background and definitions in a research workflow.
MediaWiki
Professional Wiki
A maintained MCP server that connects an agent to any MediaWiki — including Wikipedia — to search, read, create, and edit pages.
The MediaWiki server connects to any MediaWiki, including domain-specific wikis, to search, read, create, and edit pages, useful when the authoritative knowledge lives in a specialist wiki rather than Wikipedia.