OpenAlex MCP Server: 200M Academic Papers, Authors, Institutions for AI Agents

An Apify-hosted MCP server exposing 8 tools to query OpenAlex — the open catalogue of 200M+ scholarly works, authors, institutions, sources, and concepts. HTTP-only, no API key. $0.005 per tool-call.

⭐ Bookmark this OpenAlex MCP Server — Apify ranks actors by bookmarks, so it directly helps the visibility of this actor on the Apify Store. → Bookmark it on the Apify Store

What can this OpenAlex MCP Server do?

• 📚 Search 200M+ scholarly works — papers, books, datasets across every discipline.
• 👩‍🔬 Find authors with h-index, ORCID, institution — full bibliometric profiles.
• 🏛️ Resolve institutions & ROR IDs — universities, hospitals, research orgs in 200+ countries.
• 🧠 Query the OpenAlex concept taxonomy — Wikidata-linked research topics, levels 0-5.
• 📰 Discover sources — journals, conferences, repositories, OA status, DOAJ flag.
• 🤖 MCP-shaped for AI agents — call from Claude, Cursor, Continue with one tool dispatcher.

What data can you extract with the OpenAlex MCP Server?

Field	Description
openalex_id	Stable OpenAlex identifier (W… / A… / I… / S… / C…)
doi	DOI of the work (when available)
title	Title of the work / author / institution / source
publication_year	Year published
publication_date	ISO publication date
cited_by_count	Total citation count
authorships	Full nested authorship objects (raw OpenAlex)
authors	Flat list of author display names
primary_location / primary_source	Hosting journal / repo info
open_access	is_oa, oa_status, oa_url (Unpaywall-style)
concepts	Wikidata-linked research topics with score
type	Work type (article, book, dataset, …)
language	ISO 639 code
referenced_works_count	Number of references the work cites
abstract_inverted_index	Raw inverted index (word → positions)
abstract	Reconstructed plain-text abstract (when reconstruct_abstract: true)
bookmark_cta	Last record — bookmark the actor to support visibility

How to use the OpenAlex MCP Server

1. Open the actor on Apify: apify.com/dltik/mcp-server-openalex.
2. Discover the tool catalog — run once with mode=list_tools (free, no charge).
3. Pick a tool — search_works, get_work, search_authors, etc.
4. Pass arguments — e.g. {"query": "climate change", "limit": 10}.
5. Wire it into your agent — Claude / Cursor / Continue / LangChain / LlamaIndex, via Apify's MCP bridge or run-sync-get-dataset-items.

How much does the OpenAlex MCP Server cost?

PAY_PER_EVENT — $0.005 per tool-call. list_tools mode is free (discovery). Failed runs are not charged.

Plan	Tool calls	Cost
Quick	100	$0.50
Standard	1,000	$5.00
Deep	10,000	$50.00
Massive	100,000	$500.00

Per call, the OpenAlex MCP Server averages ~5 s of compute on 256 MB RAM and ~0.05 MB of bandwidth — fully covered by the per-event price.

Input parameters

Field	Type	Description
mode	enum	call_tool (charged) or list_tools (free).
tool_name	enum	search_works, get_work, search_authors, get_author, search_institutions, get_institution, search_sources, search_concepts.
arguments	object	Tool arguments — see list_tools for the per-tool schema.
mailto	string	Optional. Email passed to OpenAlex's polite-pool (~100k req/day).
reconstruct_abstract	bool	Default true — rebuilds the abstract from OpenAlex's inverted index.

Output example

{
  "openalex_id": "W2741809807",
  "doi": "https://doi.org/10.7717/peerj.4375",
  "title": "The state of OA: a large-scale analysis of the prevalence and impact of Open Access articles",
  "type": "article",
  "publication_year": 2018,
  "publication_date": "2018-02-13",
  "language": "en",
  "cited_by_count": 1132,
  "referenced_works_count": 62,
  "open_access_status": "gold",
  "open_access_url": "https://peerj.com/articles/4375.pdf",
  "is_oa": true,
  "primary_source": {
    "id": "S2738156628",
    "display_name": "PeerJ",
    "issn_l": "2167-8359",
    "type": "journal"
  },
  "authors": ["Heather Piwowar", "Jason Priem", "Vincent Larivière", "Juan Pablo Alperin"],
  "n_authors": 8,
  "concepts": [
    {"id": "C2779343474", "display_name": "Open access", "level": 2, "score": 0.94}
  ],
  "abstract": "Despite growing interest in Open Access (OA) to scholarly literature ..."
}

Use cases for the OpenAlex MCP Server

• AI research assistants — let Claude or Cursor pull verifiable citations from OpenAlex during a literature review.
• Grant writing & competitive intelligence — surface the top-cited authors and institutions on a topic in seconds.
• CRM / lead-gen enrichment for academic SaaS — enrich researcher records with h-index, ORCID, last institution, country.
• Bibliometric dashboards — track OA adoption, citation counts, journal coverage across a discipline.
• Knowledge graph builders — pull OpenAlex's concept taxonomy and author–institution edges into Neo4j / RDF.
• Systematic reviews — automate PRISMA-style filtering by year, OA status, concept score.

API examples

Python (Apify client):

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("dltik/mcp-server-openalex").call(run_input={
    "mode": "call_tool",
    "tool_name": "search_works",
    "arguments": {"query": "transformer architecture", "limit": 5, "sort": "cited_by_count:desc"}
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item.get("title"), "—", item.get("cited_by_count"))

curl (synchronous run):

curl -X POST "https://api.apify.com/v2/acts/dltik~mcp-server-openalex/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "call_tool",
    "tool_name": "get_work",
    "arguments": {"doi": "10.1038/s41586-021-03819-2"}
  }'

FAQ — OpenAlex MCP Server

Why use the OpenAlex MCP Server on Apify vs hitting api.openalex.org directly? Maintenance + agent-shape. The OpenAlex MCP Server normalizes nested objects, reconstructs abstracts from OpenAlex's inverted index, exposes a tool dispatcher native to MCP, and bills per call — no rate-limit handling, no schema-drift code in your agent.

What's the polite pool and do I need to set mailto? OpenAlex serves anonymous requests with stricter limits. Passing an email via mailto puts you in the "polite pool" with ~100k req/day. The OpenAlex MCP Server injects a maintainer email by default — set your own mailto for high-volume usage.

How are abstracts returned? OpenAlex stores abstracts as an inverted index ({word: [positions]}). When reconstruct_abstract: true (default), the OpenAlex MCP Server rebuilds the plain-text abstract into the abstract field. The raw abstract_inverted_index is always kept too.

How do I paginate beyond 10,000 results? Pass cursor: "*" on the first call, then forward the next cursor returned by OpenAlex. The OpenAlex MCP Server exposes the cursor argument on search_works for exactly this.

Is there a free tier? list_tools is always free (catalog discovery). call_tool is $0.005 per call, billed only on success.

---

⭐ Found the OpenAlex MCP Server useful? Bookmark it — it's the strongest signal for Apify Store visibility. → Bookmark on the Apify Store

Other scrapers by dltik

Actor	What it does
arXiv MCP Server	5 tools for AI agents to query arXiv preprints
HackerNews MCP Server	6 tools over 30M HN stories + comments
SEC EDGAR MCP Server	10K+ US public companies for AI agents
Pappers MCP Server	26M French companies (SIREN, dirigeants, finances)
Wikipedia MCP Server	Search, summaries, sections — all Wikipedia languages

License: MIT · Author: dltik