Paper Search MCP Server
Academic paper retrieval from multiple sources via unified search and download tools.
Data last scanned yesterday · Reviewed last month
Overview
Paper Search MCP Server provides a specialized interface for finding and downloading scholarly articles. It utilizes a two-layer architecture that combines high-level unified search and download capabilities with specific platform connectors to fetch academic materials for LLM-integrated workflows. The project emphasizes a free-first strategy, prioritizing open-access data sources while offering optional API key support to improve metadata quality and stability. It is designed for use as an MCP server with Claude Desktop or as a CLI skill within Claude Code environments.
Our verdict
With a security score of 55/100, this server is classified as high-risk. A primary security concern is that it lacks a read-only mode, meaning every installation provides full write access to the underlying system. Given the lack of formal authentication reviews and its community-maintained status, it is best suited for users comfortable with the security implications of running experimental tools in isolated environments.
- Zero direct dependencies reduce the supply-chain attack surface.
- Active development with a recent commit history.
- Supports multiple search sources with deduplication for LLMs.
- Detailed README documentation for multiple installation methods.
- Lacks a read-only mode, defaulting to full write access.
- No detected OAuth support for managing external platform permissions.
- Community-maintained without official vendor oversight.
- No homepage link provided in the repository configuration.
Tools
| Tool | Description | Risk |
|---|---|---|
| search_papers | High-level tool for multi-source concurrent search & deduplication. | low |
| download_with_fallback | Download tool relying on publisher open access links with sequential fallbacks. | low |
| search_ieee | Search IEEE Xplore. | low |
| search_acm | Search ACM Digital Library. | low |
| download_ieee | Download paper from IEEE Xplore. | low |
| download_acm | Download paper from ACM Digital Library. | low |
| read_ieee_paper | Read paper from IEEE Xplore. | low |
| read_acm_paper | Read paper from ACM Digital Library. | low |
| search_unpaywall | Standalone lookup for Unpaywall DOI-centric OA metadata. | low |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| ChatGPT | ||||
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf |
Frequently asked questions
›What are the Python version requirements for running this server?
This server requires Python 3.10 or higher.
›Are paid API keys required to use the paper search functionality?
No, the project follows a free-first strategy focusing on open and public data sources. While optional API keys can improve stability or coverage for certain sources like Semantic Scholar, the server is designed to be usable without them.
›Can I use this server with Claude Desktop?
Yes, it is compatible with MCP clients including Claude Desktop and can be installed via methods such as Smithery or standard package managers like pip and uv.
›Does the server support direct full-text reading for all academic platforms?
No, platform capabilities vary. While sources like arXiv and bioRxiv support search, download, and reading, others like PubMed currently only provide information, and Google Scholar is restricted by bot-detection.
›What is the role of the download_with_fallback tool?
This tool attempts to retrieve papers by following an open-access priority chain, starting with source-native downloads and falling back through various open repositories and Unpaywall resolution before reaching optional Sci-Hub workflows.
Alternatives
Badge
Maintain this server? Add the live badge to your README.