PubMed Scraper: medical research articles, abstracts and MeSH terms

Export articles from PubMed, the US National Library of Medicine database of 38 million+ biomedical citations. Enter any PubMed search and get clean, flat rows ready for Excel, Google Sheets, BI tools, your database or an LLM/RAG pipeline. Each row has the abstract, DOI, PMCID, journal, publication date, authors, affiliations, MeSH terms, keywords and publication types.

Use it as a no-code PubMed API. It runs on the official NCBI E-utilities API, so you need no proxies and no browser, and it costs $1.50 per 1,000 articles.

What does the PubMed Scraper do?

It turns a PubMed search into a dataset with one row per article. Each row includes:

• PMID, DOI, PMCID (when the article is in PubMed Central) and the PubMed link
• title, and the full abstract with its section labels (BACKGROUND, METHODS, RESULTS...)
• journal name, ISO journal abbreviation, publication date and year
• all authors, author count, the first author's affiliation and every unique affiliation
• MeSH terms, author keywords and publication types (Review, Randomized Controlled Trial, Meta-Analysis...)

Who is it for?

• Pharma, biotech and medical affairs. Monitor new publications on a drug, target or competitor. Find key opinion leaders (KOLs) and their institutions from author affiliations.
• Researchers and librarians. Build datasets for systematic reviews, bibliometrics and meta-research. Filter by publication type or MeSH term in your spreadsheet.
• AI and data teams. Feed abstracts into RAG pipelines, embeddings, classifiers or LLM summarizers. The rows are flat and already clean text.
• Medtech sales and recruiters. Find which hospitals and labs publish on a topic and who leads the work.
• Monitoring and alerts. Schedule a daily run with the date type set to Entrez date to get only the articles newly added to PubMed.

Why use this PubMed scraper?

• Low price. $1.50 per 1,000 articles. You pay only for rows saved, and you can set a maximum charge per run.
• Richer rows. DOI, PMCID, labeled abstract sections, affiliations, MeSH terms, keywords and publication types in every row, not just title and authors.
• Official API, not HTML scraping. It runs on NCBI E-utilities, fetching 200 articles per request. It doesn't break when the website changes.
• Full PubMed syntax. Field tags ([Title], [MeSH Terms], [Author], [Journal], [dp]), Boolean operators and phrases all work exactly like on pubmed.ncbi.nlm.nih.gov.
• Polite and reliable. It stays within NCBI rate limits (3 requests per second, or 10 with your free API key) and retries on errors.
• Fast and light. The default 50-article run finishes in a few seconds on 256 MB of memory.

How to scrape PubMed

1. Click Try for free (or Start) on this page.
2. Type a PubMed query, such as semaglutide AND obesity.
3. Optionally set a date range, a sort order and Max articles (this also caps your cost).
4. Click Start and wait a few seconds.
5. Download the results as CSV, Excel, JSON, XML or HTML, or read them via the API.

Input examples

Most recent randomized controlled trials on semaglutide published in 2025:

{
  "query": "semaglutide AND randomized controlled trial[pt]",
  "minDate": "2025-01-01",
  "maxDate": "2025-12-31",
  "dateType": "pdat",
  "sort": "pub_date",
  "maxResults": 1000
}

Daily monitoring: everything added to PubMed yesterday that mentions CRISPR in the title, without abstracts:

{
  "query": "CRISPR[Title]",
  "minDate": "2026-10-22",
  "maxDate": "2026-10-22",
  "dateType": "edat",
  "includeAbstract": false,
  "maxResults": 500
}

Field	What it does	Example
query	PubMed search, with full PubMed syntax	CRISPR gene therapy, Smith J[Author] AND cancer
minDate / maxDate	Date range. Format YYYY, YYYY-MM or YYYY-MM-DD	2024, 2025-06-30
dateType	pdat = publication date, edat = date added to PubMed	edat
sort	relevance (best match) or pub_date (most recent first)	pub_date
maxResults	Maximum articles to save (default 50, max 10,000)	1000
includeAbstract	Include the abstract text (default true)	false
apiKey	Optional free NCBI API key for faster large exports (stored as a secret)

Output sample

{
  "pmid": "40211735",
  "doi": "10.1097/CM9.0000000000003567",
  "pmcid": "PMC13186126",
  "title": "Epigenetically elevated RAD51AP1 regulates the RAD51-UAF1 complex contributing to temozolomide resistance in EGFRvIII glioblastoma.",
  "abstract": "BACKGROUND: Epidermal growth factor receptor variant III ( EGFRvIII ) mutation is the most prevalent genetic change in glioblastoma...\nMETHODS: We constructed a Clustered regularly interspaced shortpalindromic repeats (CRISPR)/CRISPR-associated (Cas) system 9 library...\nRESULTS: ...\nCONCLUSION: ...",
  "journal": "Chinese medical journal",
  "journalIsoAbbreviation": "Chin Med J (Engl)",
  "pubDate": "2026-05-20",
  "year": 2026,
  "authors": "Chang Y, Yang J, Li Z, Zhu X, Jiang T, Wang Q, Yu M, Han B, Liu X",
  "authorCount": 9,
  "firstAuthorAffiliation": "Beijing Neurosurgical Institute, Capital Medical University, Beijing 100070, China.",
  "affiliations": "Beijing Neurosurgical Institute, Capital Medical University, Beijing 100070, China.; Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing 100070, China.; Department of Neurosurgery, Tianjin Medical University General Hospital, Tianjin 300052, China.",
  "meshTerms": "Glioblastoma; Temozolomide; Humans; Rad51 Recombinase; ErbB Receptors; Cell Line, Tumor; Drug Resistance, Neoplasm; ...",
  "keywords": "Epidermal growth factor receptor variant III; Glioblastoma; Temozolomide resistance",
  "publicationTypes": "Journal Article",
  "url": "https://pubmed.ncbi.nlm.nih.gov/40211735/",
  "scrapedAt": "2026-10-23T08:12:41.120Z"
}

Field	Type	Description
pmid	string	PubMed identifier
doi	string	Digital Object Identifier, when available
pmcid	string	PubMed Central ID, when the full text is in PMC (often open access)
title	string	Article title (inline formatting removed)
abstract	string	Abstract; structured sections are labeled and separated by line breaks
journal, journalIsoAbbreviation	string	Journal name and its ISO abbreviation
pubDate, year	string, number	Publication date (YYYY, YYYY-MM or YYYY-MM-DD; sometimes a range like 2019 Nov-Dec) and year
authors, authorCount	string, number	Authors as LastName Initials, comma-separated, and their number
firstAuthorAffiliation	string	First author's first affiliation
affiliations	string	All unique affiliations, separated by ;
meshTerms	string	MeSH descriptors, separated by ;
keywords	string	Author keywords, separated by ;
publicationTypes	string	Publication types, separated by ;
url	string	Link to the PubMed page
scrapedAt	string	When the row was collected (ISO 8601)

How much does it cost to scrape PubMed?

This Actor uses pay-per-event pricing:

• $0.0015 per article saved ($1.50 per 1,000 articles)
• $0.00005 per run start (per GB of memory, a fraction of a cent)

Worked example: exporting 5,000 articles on a topic costs 5,000 × $0.0015 = $7.50, plus $0.00005 × 0.25 GB ≈ $0.00001 for the run start. The default 50-article run costs about $0.08, so the $5 of monthly free credit on the Apify Free plan covers over 3,000 articles.

Set Max articles or a maximum charge per run, and the Actor stops as soon as it reaches the limit.

Integrations and API

Call the Actor from your own code with the official apify-client, or connect it to Make, Zapier, n8n, Google Sheets, Slack, webhooks and more from the Integrations tab.

Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("deanko/pubmed-scraper").call(run_input={
    "query": "long covid AND fatigue",
    "minDate": "2025",
    "sort": "pub_date",
    "maxResults": 500,
})
for article in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(article["pmid"], article["doi"], article["title"])

JavaScript / Node.js

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('deanko/pubmed-scraper').call({
    query: 'Alzheimer disease[MeSH Terms] AND lecanemab',
    maxResults: 200,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.length, 'articles');

You can also run it with one HTTP request through the run-sync-get-dataset-items endpoint and get the rows back in the response.

FAQ

Where does the data come from? It comes from the official, public NCBI E-utilities API run by the US National Library of Medicine (NIH). The data is the same as on pubmed.ncbi.nlm.nih.gov.

Do I need an NCBI API key? No. Without a key the Actor makes at most 3 requests per second, which is about 600 articles per second of fetching anyway. A free key (from your NCBI account settings) raises the limit to 10 requests per second and helps with very large exports. It is stored as a secret input.

How many articles can I get from one search? Up to 10,000, which is a PubMed limit for any single search. For more, split your query into date ranges (for example one run per year) and run them one after another.

Can I use the same search syntax as on the PubMed website? Yes. Copy the query from the PubMed search box, including field tags such as [Title/Abstract], [MeSH Terms], [Author], [Journal] and [pt]. The run log shows how PubMed translated it.

Why do some results fall outside my publication date range? PubMed's publication date covers both the print and the electronic date. An article published online in 2025 can carry a 2026 print issue date, and pubDate shows the journal issue date. Use edat to filter by the date the article was added to PubMed.

Why is the abstract empty for some articles? Some records, such as letters, editorials, comments and older citations, have no abstract in PubMed.

Do I get the full text? No, PubMed holds citations and abstracts. When pmcid is filled, the full text is available in PubMed Central.

Limitations

• Only PubMed citations are returned. Book chapters (NCBI Bookshelf records) are skipped.
• Full text, references, grant numbers and author ORCIDs are not included.
• One search returns at most 10,000 articles (a PubMed limit).
• The data is only as fresh as PubMed. New citations usually appear within a day of publication.

Legal and privacy

The data comes from the public PubMed database through the official NCBI E-utilities API. This Actor is not affiliated with or endorsed by the US National Library of Medicine, NCBI or NIH. Abstracts may be protected by publishers' copyright. You are responsible for using the data in line with the NLM terms and the laws that apply to you.

Support and feedback

Missing a field or filter you need, or found a bug? Open an issue in the Issues tab on this page. If the Actor saves you time, a short review helps other people find it.