paper-fetch

Paper Fetch

Overview

Searches multiple academic databases and downloads user-selected papers as PDFs. Uses paper-search-mcp for multi-source concurrent search with deduplication and open-access fallback.

Required MCP

openags/paper-search-mcp must be installed and configured.

Install:

npx -y @smithery/cli install @openags/paper-search-mcp --client claude

Input

Option A — Keyword Search

Ask the user for:

• Search terms (required)
• Sources to search: arxiv, pubmed, semantic_scholar, biorxiv (default: all)
• Date range, e.g. 2022-2025 (optional)
• Max results, e.g. 10 (default: 10)

Option B — DOI / URL List

Accept a plain list, one per line:

10.48550/arXiv.1706.03762
10.1038/s41586-021-03819-2
https://arxiv.org/abs/2301.00001

Steps

1. Call search_papers (Option A) or skip directly to step 3 (Option B)

2. Present candidate table to user:

   #	Title	Authors	Year	Source
   1	Attention Is All You Need	Vaswani et al.	2017	arxiv

   Include one-line abstract excerpt under each row.

3. Ask: "Which papers to download? (e.g. 1,3,5 or 'all')"

4. Call download_with_fallback for each selected paper (tries publisher link → Unpaywall → open access)

5. Return structured output (see below)

Output

Pass this JSON to the next step (paper-zotero or paper-note):

{
  "papers": [
    {
      "title": "Attention Is All You Need",
      "authors": ["Vaswani", "Shazeer", "Parmar"],
      "year": 2017,
      "doi": "10.48550/arXiv.1706.03762",
      "source": "arxiv",
      "pdf_path": "/tmp/papers/attention-is-all-you-need.pdf",
      "abstract": "We propose a new simple network architecture..."
    }
  ]
}

Common Issues

• PDF unavailable: download_with_fallback tries Unpaywall automatically. If still unavailable, set "pdf_path": null and continue with remaining papers.
• Too many results: Remind user to narrow date range or add more specific terms.