PubMed Articles Scraper — Authors, Journals & DOIs

Search PubMed and get one row per article: title, the full author list, journal, volume/issue/pages, a bare DOI with a resolvable link, PMCID, publication types and languages.

Full PubMed query syntax. No API key, no account. $0.008 per article — a search that matches nothing is recorded and not charged.

Input

{ "query": "crispr AND 2024[pdat]", "maxArticles": 100, "sort": "date" }

Bulk:

{ "queries": ["semaglutide", "\"Smith J\"[au] AND cardiology"], "maxArticles": 50 }

Field	Type	Notes
query	string	Any PubMed query — crispr AND 2024[pdat], "Smith J"[au], MeSH terms.
queries	string[]	Up to 10 searches per run.
maxArticles	integer	1–500 per search. You are charged per article returned.
sort	enum	relevance, date, author, journal.
apiKey	string	Optional NCBI key — see Limits.

Output

{
  "ok": true,
  "pmid": "25315507",
  "url": "https://pubmed.ncbi.nlm.nih.gov/25315507/",
  "title": "Genome modification by CRISPR/Cas9.",
  "authors": ["Ma Y", "Zhang L", "Huang X"],
  "firstAuthor": "Ma Y",
  "journal": "The FEBS journal",
  "pubDateRaw": "2014 Dec",
  "pubDate": null,
  "pubYear": 2014,
  "volume": "281", "issue": "23", "pages": "5186-93",
  "doi": "10.1111/febs.13110",
  "doiUrl": "https://doi.org/10.1111/febs.13110",
  "pmcid": null,
  "publicationTypes": ["Journal Article", "Review"],
  "languages": ["eng"],
  "totalMatches": 70754
}

What it gets right

• ⭐ Partial dates are not invented. PubMed prints "2014 Dec" when the issue was monthly. pubDateRaw keeps exactly what the source said, pubDate stays null unless a full day exists, and pubYear is always there for filtering. No article gets a fake 1st-of-the-month.
• The DOI is bare and the link works. PubMed returns "doi: 10.1111/…" in elocationid; the prefix is stripped and doiUrl is built from the cleaned value.
• ⚠️ esummary returns a map keyed by id plus a uids array. Iterating its values naively emits that array as a bogus, title-less record — which you would then be billed for. This walks the ids it asked for, in order, and the promise spec asserts every pmid is numeric.
• totalMatches tells you how big the literature is, so you know whether 25 rows is the whole answer or the tip of 70,000.
• Two calls, batched. esearch for ids, then esummary in batches of 200 — not one request per article.

Pricing

$0.008 per article returned. No start fee. A search matching nothing produces an articleCount: 0 record with totalMatches and is not charged.

Limits

• NCBI allows 3 requests/second per IP without a key and 10/second with one. Runs here are small and batched, so the keyless limit is usually fine; supply apiKey (free from your NCBI account) for heavy use.
• 10 searches per run, up to 500 articles each.
• Returns citation metadata, not abstracts or full text. Use the pmcid for open-access full text via PubMed Central.
• citedByCount is PubMed Central's reference count where available, not a Scopus/WoS citation count.

FAQ

Do I need an NCBI API key? No.

Can I get abstracts? Not in this Actor — it returns citation metadata.

Does PubMed query syntax work? Yes, including field tags ([au], [pdat], [mesh]) and boolean operators.

Changelog

• 0.1 (2026-09-10) — first release. One row per article, batched esummary, bare DOIs with resolvable links, partial dates preserved, bulk searches.