A
licenseNot graded
qualityD
maintenanceEnables searching and retrieving academic paper metadata from Semantic Scholar, including paper details, citations, and author information.
20
MIT
paper-mcp
by imnotdev25
This server lets you search, retrieve, and analyze academic papers, authors, citations, and bibliographies through MCP tools.
Paper metadata: get title, authors, abstract, DOI, arXiv ID, citation count, TL;DR, open-access status, fields of study.
PDF access: find the best open-access PDF URL for a paper.
Full text: retrieve full text, structured sections, and extracted tables when available.
Citations & references: list up to 100 papers citing a paper or cited by it.
Search: run filtered, deduplicated searches across Semantic Scholar, OpenAlex, PubMed, and DBLP.
DOI resolution: resolve a DOI to merged metadata from Crossref, Semantic Scholar, and Unpaywall.
Author tools: search authors, view profiles (h-index, affiliations, counts), and list a researcher's papers.
BibTeX: generate BibTeX citations from Semantic Scholar or Crossref.
Similar papers: get related-paper recommendations.
Batch lookup: look up up to 500 paper IDs in one request.
Provides tools to retrieve academic papers from arXiv, including preprint information and full text.
Provides tools to retrieve academic paper metadata, citations, and references from Semantic Scholar.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@paper-mcp Get metadata for 'Attention Is All You Need'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
📄 paper-mcp
An MCP server built with FastMCP for paper, citation, author, and bibliography search. Run it in one command with uvx — no manual install needed.
✨ Features
Paper tools
Tool | Returns |
| Title, authors, abstract, DOI, arXiv ID, citation count, TL;DR, OA status, fields of study |
| Best open-access PDF URL |
| Full text plus structured pages, sections, and extracted tables |
| Up to 100 papers that cite this one |
| Up to 100 papers this one cites |
| Filtered, deduplicated search across S2, OpenAlex, PubMed, and DBLP |
| Venue, publisher, version of record, and preprint status |
| BibTeX citation from Semantic Scholar or Crossref |
| Related-paper recommendations |
| Up to 500 paper IDs in one Semantic Scholar request |
| DOI metadata merged from Crossref, Semantic Scholar, and Unpaywall |
Author and source tools
Tool | Returns |
| Disambiguated Semantic Scholar author matches |
| h-index, affiliations, paper and citation counts |
| A researcher's bibliography |
| Broad OpenAlex work search with institutions and topics |
| PubMed biomedical article summaries |
| Curated DBLP computer-science bibliography results |
paper_search accepts sources, year range, venue, author, publication type,
minimum citations, open-access-only, and relevance/year/citation sorting.
Full text uses arXiv HTML → structured open-access PDF extraction → Lightpanda → abstract.
Responses from repeated Semantic Scholar searches and batch lookups are cached in memory.
Related MCP server: Academic Paper Search MCP Server
🚀 Quick Start
Run without installing (uvx)
# stdio mode — for Claude Desktop / most MCP clients
uvx paper-mcp
# SSE mode — for remote or multi-client setups
uvx paper-mcp --transport sse --port 8000
uvxdownloads, installs (in an isolated env), and runs the package — zero setup.
Install permanently
uv tool install paper-mcp
paper-mcp # now available globally
paper-mcp --transport sseLocal development
git clone https://github.com/imnotdev25/paper-search
cd paper-search
uv sync # install all deps from pyproject.toml
uv run paper-mcp # run directly
uv run paper-mcp --transport sse🖥 Claude Desktop Config
Add to claude_desktop_config.json:
{
"mcpServers": {
"papers": {
"command": "uvx",
"args": ["paper-mcp"]
}
}
}No Python paths, no venv activation — uvx handles everything.
🌐 Browser Fallback (gomcp / Lightpanda)
For JS-rendered publisher pages, the server automatically starts a Lightpanda headless browser via gomcp.
One-time setup:
# Download gomcp binary from GitHub releases:
# https://github.com/lightpanda-io/gomcp/releases
# Then download the Lightpanda browser binary:
gomcp downloadIf gomcp is not installed, the server still works — browser-dependent
paths fall back to abstract/metadata gracefully.
🏗 Architecture
Claude (LLM)
│ MCP (stdio or SSE)
▼
paper-mcp [FastMCP, Python]
│
├── Semantic Scholar API ── metadata, citations, references
├── arXiv API + HTML ── preprint info + full text
├── Crossref + Unpaywall ── DOI metadata, BibTeX, OA PDFs
├── OpenAlex / PubMed / DBLP ── corpus-specific search
└── gomcp SSE ───────────── Lightpanda browser (JS fallback)
│ CDP
└── Lightpanda Browser (headless)📦 Publishing to PyPI
# Build
uv build
# Publish (needs PyPI token)
uv publish --token $PYPI_TOKENOnce on PyPI, anyone can run it with uvx paper-mcp.
⚙️ CLI Options
usage: paper-mcp [-h] [--transport {stdio,sse}] [--port PORT] [--host HOST]
options:
--transport stdio (default) or sse
--port SSE port (default: 8000)
--host SSE host (default: 127.0.0.1)🔑 Notes
Set
S2_API_KEYfor authenticated Semantic Scholar requests.Set
UNPAYWALL_EMAILto a real contact address to enable Unpaywall. Requests are skipped when it is unset; no placeholder identity is sent.Optionally set
OPENALEX_EMAILfor the OpenAlex polite pool.PAPER_CACHE_TTLcontrols the in-memory cache in seconds (default3600;0disables it).PDF extraction is limited to open PDFs up to 25 MB and 50,000 extracted characters. Paywalled PDFs still require institutional access.
📁 Project Structure
paper-mcp/
├── pyproject.toml ← packaging, entry point, deps
├── README.md
├── src/
└── paper_mcp/
├── __init__.py
└── server.py ← FastMCP tools + main()
└── tests/
└── test_server.py ← focused HTTP, PDF, cache, and config testsAvailable Tools
6 toolsdoi_get_metadataResolve DOI to Full Paper MetadataARead-onlyIdempotent
Resolve a DOI to comprehensive paper metadata from multiple authoritative sources.
This is the reverse of paper_get_metadata: given a DOI you already know, retrieve everything about the paper without needing its title.
Data is merged from three complementary sources (all queried in parallel):
Crossref — canonical DOI registry; best for bibliographic data, publisher, volume/issue/page, funders, license.
Semantic Scholar — adds citation count, references, fields of study, TL;DR, and open-access PDF link.
Unpaywall — adds detailed open-access status and every known PDF URL.
Returns a merged JSON object with: doi, title, abstract, authors, year, publication_date, venue, publisher, volume, issue, page, type, subjects, issn, funders, license, citation_count, reference_count, is_referenced_by_count, fields_of_study, tldr, publication_types, is_open_access, oa_status, open_access_pdf, oa_locations (list of all known PDF URLs with host_type / version / license), arxiv_id, semantic_scholar_id, sources (list of which APIs responded successfully)
Args: params (DoiInput): { doi: str } Accepts bare DOI or full URL, e.g.: "10.1145/3442188.3445922" "https://doi.org/10.1145/3442188.3445922"
Returns: str: JSON-encoded metadata dict or { "error": "..." }.
Inputs: params (required). Returns: result
paper_get_citationsGet Papers That Cite This PaperARead-onlyIdempotent
Return up to 100 papers that cite the given paper (forward citations).
Returns JSON with: paper_title, total_citations, returned, citations (list of {title, authors, year, venue, doi, arxiv_id, citation_count}).
Args: params (PaperInput): { paper_title: str }
Returns: str: JSON-encoded citations or { "error": "..." }.
Inputs: params (required). Returns: result
paper_get_fulltextGet Paper Full TextARead-onlyIdempotent
Retrieve the full text of a paper by title (up to 50,000 characters).
Fetch order: arXiv HTML5 renderer → gomcp browser → abstract-only fallback.
Returns JSON with: title, abstract, tldr, fulltext (str|null), fulltext_length (int), source.
Args: params (PaperInput): { paper_title: str }
Returns: str: JSON-encoded result or { "error": "..." }.
Inputs: params (required). Returns: result
paper_get_metadataGet Paper MetadataARead-onlyIdempotent
Return comprehensive metadata for an academic paper searched by title.
Queries Semantic Scholar first (richest metadata), then falls back to arXiv.
Returns JSON with: title, authors, abstract, year, venue, doi, arxiv_id, semantic_scholar_id, citation_count, reference_count, fields_of_study, publication_types, publication_date, is_open_access, open_access_pdf, tldr.
Args: params (PaperInput): { paper_title: str }
Returns: str: JSON-encoded metadata dict or { "error": "..." }.
Inputs: params (required). Returns: result
paper_get_pdfGet Paper PDF URLARead-onlyIdempotent
Find the best open-access PDF URL for a paper by title.
Source priority: Semantic Scholar OA → arXiv PDF → Unpaywall → gomcp browser.
Returns JSON with: pdf_url, source, title, year. May include browser_excerpt if gomcp fallback was used.
Args: params (PaperInput): { paper_title: str }
Returns: str: JSON-encoded result or { "error": "..." }.
Inputs: params (required). Returns: result
paper_get_referencesGet Paper Reference ListARead-onlyIdempotent
Return the bibliography of a paper — the papers it cites (backward citations).
Returns JSON with: paper_title, total_references, returned, references (list of {title, authors, year, venue, doi, arxiv_id}).
Args: params (PaperInput): { paper_title: str }
Returns: str: JSON-encoded references or { "error": "..." }.
Inputs: params (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
A4.2/5.0
Scored across 6 tools
Disambiguation5/5
Each tool has a clearly distinct purpose: metadata by DOI, metadata by title, forward citations, full text, PDF URL, and backward citations. The slight overlap between doi_get_metadata and paper_get_metadata is clarified by their different inputs (DOI vs title).
Naming Consistency3/5
Tool names are somewhat inconsistent: most use 'paper_get_' prefix, but one uses 'doi_get_' instead. While all are readable, the mix of prefixes and the inversion in 'doi_get_metadata' compared to 'paper_get_metadata' breaks the otherwise verb_noun pattern.
Tool Count5/5
With 6 tools, the server is well-scoped for its purpose. Each tool provides a distinct function without unnecessary bloat, and the count is within the ideal 3-15 range.
Completeness4/5
The tool set covers key operations for a paper metadata server: metadata retrieval (by title and DOI), full text, PDF URL, citations, and references. Minor gaps like search by author or keyword are absent, but the core workflow is supported.
Maintenance
ActivityInactive
ResponsivenessNo recent community issues
Related MCP Connectors
Find academic papers across major sources like arXiv, PubMed, bioRxiv, and more. Download PDFs whe…
Search and download academic papers from arXiv, PubMed, bioRxiv, medRxiv, Google Scholar, Semantic…
Search arXiv/Semantic Scholar/OpenAlex + medical evidence (PubMed/Europe PMC) + LaTeX/PDF tools.
Find research papers and arXiv preprints, and get an abstract and free full text from a DOI or id.
Related MCP Servers
- FlicenseNot gradedqualityAmaintenanceEnables searching and retrieving academic paper information from Semantic Scholar and Crossref, including metadata, abstracts, full-text when available, and structured responses for LLM tool use.-
- FlicenseNot gradedqualityAmaintenanceEnables searching and retrieving academic paper information from Semantic Scholar and Crossref, including metadata, abstracts, full-text when available, and topic/date-range searches.-
- FlicenseNot gradedqualityAmaintenanceEnables searching for academic papers, authors, and citation networks through the Semantic Scholar API. Supports full-text and title search, batch lookups, citation and reference tracking, author profiles, and paper recommendations.-