papers-mcp
An MCP server that exposes the OpenAlex academic research database to AI assistants. 240M+ scholarly works, authors, journals, institutions, topics, publishers, and funders — all queryable from Claude.
Tools
| Tool | Description |
|---|---|
work_list |
Search and filter scholarly works |
work_get |
Get a single work by ID, DOI, PMID, or PMCID |
work_autocomplete |
Type-ahead search for works by title |
work_find |
AI semantic search (requires API key) |
author_list |
Search and filter researcher profiles |
author_get |
Get a single author |
author_autocomplete |
Type-ahead search for authors |
source_list |
Search journals, repositories, conferences |
source_get |
Get a single source |
source_autocomplete |
Type-ahead search for sources |
institution_list |
Search universities, hospitals, companies |
institution_get |
Get a single institution |
institution_autocomplete |
Type-ahead search for institutions |
topic_list |
Search the topic hierarchy |
topic_get |
Get a single topic |
publisher_list |
Search publishing organizations |
publisher_get |
Get a single publisher |
publisher_autocomplete |
Type-ahead search for publishers |
funder_list |
Search grant-making organizations |
funder_get |
Get a single funder |
funder_autocomplete |
Type-ahead search for funders |
Setup
Build
The binary is at target/release/papers-mcp (or papers-mcp.exe on Windows).
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json
(macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
For semantic search (work_find), add your OpenAlex API key:
Claude Code
Add to your project's .mcp.json or run:
Or configure in ~/.claude/settings.json:
Usage
Once connected, ask Claude naturally:
- "Find recent open-access papers on transformer attention mechanisms"
- "Who is Yoshua Bengio and what are his most cited works?"
- "What journals publish the most machine learning research?"
- "Show me papers by authors at MIT published in 2023"
- "Get the full details for DOI 10.7717/peerj.4375"
List response format
List tools return a slim subset of fields to keep responses concise — large arrays like
referenced_works and counts_by_year are omitted. Use the corresponding _get tool
to retrieve the full entity record.
See ../papers/CHANGES.md for the complete list of differences
from the raw OpenAlex API response.
API key
An API key is optional for most tools but enables higher rate limits and is required for
work_find (semantic search). Get one at openalex.org.