Paper Search MCP Server

Content

Paper Search MCP Server is a Python MCP server and CLI for academic literature workflows. It can search papers, deduplicate results across sources, download open-access PDFs where available, and extract text so an MCP client can work with paper metadata and content.

The project is broader than a single-source server: the README describes connectors for arXiv, PubMed, bioRxiv, medRxiv, Semantic Scholar, Crossref, OpenAlex, PubMed Central, CORE, Europe PMC, dblp, OpenAIRE, CiteSeerX, DOAJ, BASE, Zenodo, HAL, SSRN, Unpaywall, and optional restricted or paid-platform connectors. It also includes a Claude Code skill path for users who prefer a skill plus CLI workflow instead of an MCP server.

Source Review

• https://github.com/openags/paper-search-mcp
• https://github.com/openags/paper-search-mcp/blob/main/README.md
• https://github.com/openags/paper-search-mcp/blob/main/LICENSE
• https://github.com/openags/paper-search-mcp/blob/main/pyproject.toml
• https://github.com/openags/paper-search-mcp/blob/main/.env.example
• https://github.com/openags/paper-search-mcp/blob/main/Dockerfile
• https://github.com/openags/paper-search-mcp/blob/main/smithery.yaml
• https://github.com/openags/paper-search-mcp/blob/main/claude-code/SKILL.md
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/server.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/cli.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/config.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/paper.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/arxiv.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/pubmed.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/openalex.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/semantic.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/core.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/unpaywall.py
• https://github.com/openags/paper-search-mcp/blob/main/paper_search_mcp/academic_platforms/sci_hub.py
• https://pypi.org/pypi/paper-search-mcp/json

These sources were reviewed on 2026-06-05. Prefer the live repository, README, license, Python package metadata, environment example, Dockerfile, Smithery config, Claude Code skill, server and CLI modules, config model, paper model, representative connectors, and PyPI metadata for current installation, source support, and caveats.

Features

• Search papers across multiple academic sources.
• Deduplicate and standardize search results.
• Download open-access PDFs through source-native and fallback paths.
• Extract readable paper text for MCP workflows.
• Use optional API keys for Unpaywall, CORE, Semantic Scholar, DOAJ, Zenodo, and other supported services.
• Use a Claude Code skill plus CLI workflow as an alternative to MCP configuration.
• Run through uvx, uv tool install, pip, Smithery, Docker, or source checkout.
• Keep optional restricted or unstable connectors separate from the open-access-first path.

Installation

Run the MCP server with uvx:

uvx paper-search-mcp

For Claude Desktop-style JSON configuration:

{
  "mcpServers": {
    "paper-search-mcp": {
      "command": "uvx",
      "args": ["paper-search-mcp"],
      "env": {
        "PAPER_SEARCH_MCP_UNPAYWALL_EMAIL": "YOUR_EMAIL",
        "PAPER_SEARCH_MCP_SEMANTIC_SCHOLAR_API_KEY": "OPTIONAL_API_KEY",
        "PAPER_SEARCH_MCP_CORE_API_KEY": "OPTIONAL_API_KEY"
      }
    }
  }
}

Leave optional keys unset unless you need the corresponding provider and have permission to use it.

Use Cases

• Search recent papers across several academic indexes.
• Find open-access PDFs for a DOI or title.
• Read paper text into a literature-review workflow.
• Compare arXiv, PubMed, Semantic Scholar, Crossref, and OpenAlex metadata.
• Use Claude Code skills to trigger paper search without MCP server setup.
• Build a free-first research assistant for public academic sources.
• Keep restricted connectors disabled unless policy and credentials allow them.

Safety and Privacy

Paper search and download workflows can run into copyright, license, rate-limit, and institutional-access constraints. Prefer open-access sources, keep optional restricted connectors disabled by default, and treat Sci-Hub as a user-enabled fallback with jurisdiction and policy risk rather than a recommended path.

Protect source credentials and research topics. Queries, DOI lists, downloaded PDFs, extracted text, abstracts, API keys, proxy URLs, and emails can reveal private research interests or restricted content. Keep sensitive literature review work out of shared prompts and logs unless approved.

Duplicate Check

No openags/paper-search-mcp entry, Paper Search MCP entry, or matching source URL was found in content/mcp. The existing arXiv MCP entry covers a narrower single-source arXiv server.