PubMed Abstract Scraper

Extract structured PubMed abstracts by keyword — with optional date filtering, DOI links, and author lists.

Search PubMed's 35+ million biomedical citations and retrieve clean, structured JSON output: title, authors, DOI, abstract text, journal name, and publication date. No API key required.

---

What does this Actor do?

This Actor queries the NCBI PubMed database using the official E-utilities API. Given a search keyword (supports MeSH terms, gene names, author filters, and free-text queries), it retrieves abstracts and returns them as structured JSON records — one object per article.

Ideal for systematic literature reviews, research trend analysis, citation monitoring, and building medical knowledge bases.

---

Use Cases

1. Systematic Literature Review Automation

A researcher studying COVID-19 long-term effects needs 500 recent abstracts. Input: keyword: "long COVID symptoms", dateFrom: "2021-01-01", maxResults: 500. The Actor returns all matching abstracts in minutes instead of hours of manual PubMed browsing — ready to import into Zotero, Rayyan, or Excel.

2. Competitor & Drug Pipeline Monitoring

A pharma team tracks new publications about a competitor's drug every week. They run the Actor with keyword: "semaglutide cardiovascular" on a schedule and feed the output into a dashboard to detect new trial results, safety signals, or opinion pieces the moment they are published.

3. Medical Education Content Generation

A medical education platform wants to keep its question bank up to date. The Actor scrapes the latest guidelines (e.g. keyword: "ESC heart failure guidelines 2024") and the abstracts are fed to an LLM pipeline that generates new MCQ questions — ensuring currency without manual curation.

---

Input

Field	Type	Required	Default	Description
keyword	String	✅ Yes	—	PubMed search query (MeSH, gene, free text, etc.)
maxResults	Integer	No	100	Max abstracts to retrieve (1–10,000)
dateFrom	String	No	—	Filter: published on/after this date (YYYY-MM-DD)
dateTo	String	No	today	Filter: published on/before this date (YYYY-MM-DD)

Example Input

{
  "keyword": "myocardial infarction reperfusion therapy",
  "maxResults": 50,
  "dateFrom": "2022-01-01",
  "dateTo": "2024-12-31"
}

---

Output

Each result is saved to the Default Dataset as a JSON object:

{
  "pmid": "38123456",
  "title": "Outcomes of primary PCI vs thrombolysis in STEMI: a meta-analysis",
  "authors": ["Müller A", "Schmidt B", "Jensen C"],
  "doi": "10.1016/j.jacc.2023.11.042",
  "abstract": "Background: Primary percutaneous coronary intervention (PCI) is the standard of care for ST-elevation myocardial infarction... Conclusions: Primary PCI significantly reduces 30-day mortality compared to thrombolysis (OR 0.63, 95% CI 0.54–0.74).",
  "pubDate": "2024-01-15",
  "journal": "Journal of the American College of Cardiology"
}

---

Pricing

This Actor uses Pay-Per-Result pricing at $0.008 per scraped abstract.

Volume	Estimated Cost
100 abstracts	~$0.80
1,000 abstracts	~$8.00
10,000 abstracts	~$80.00

---

Technical Details

• Data source: NCBI E-utilities API (official, public, no key required)
• Rate limiting: 1 request/second (conservative; NCBI allows 3/sec without key)
• Retry logic: Up to 3 automatic retries with exponential backoff
• Batch size: 20 articles per API call (NCBI limit)
• Output format: Structured JSON, one object per abstract
• Node.js: v22 LTS

---

Supported Query Syntax

PubMed supports advanced search syntax. Examples:

• BRCA1[gene] AND cancer — gene name filter
• Smith J[author] AND cardiology — author filter
• "heart failure"[MeSH] AND randomized controlled trial[pt] — MeSH + publication type
• aspirin AND (myocardial infarction OR stroke) — boolean logic

---

Support

Issues or feature requests? Open a ticket via the Apify console.