# PubMed search: `POST /v1/pubmed/search`

Cost: 2 credits (2 per page (an empty page and failed requests are free)). Docs: https://scrapercompany.com/docs/reference/pubmed-search

One page of full PubMed records for a query, in PubMed's own ranking.

Each article carries the title, the labelled
abstract (`abstract_sections`, plus the joined `abstract`), authors
with affiliations and ORCID, the journal block, `doi`/`pmcid`,
publication types, MeSH terms and the canonical article `link`.

`term` is PubMed's query language verbatim (field tags, MeSH,
booleans); `query_translation` in the answer is PubMed's own
expansion of it. `total` is the full match count; `next_page` is set
while more remain. An empty result is a normal empty page and is not
charged.

Request body (JSON):

- `term` (string; required; min length 1, max length 1000): PubMed query, in PubMed's own syntax: words, phrases, field tags (`cancer[Title]`, `"cold water"[MeSH]`), and AND/OR/NOT all pass through verbatim.
- `page` (integer; optional; default 1; min 1, max 1000): Result page, from 1.
- `page_size` (integer; optional; default 10; min 1, max 100): Records per page; each is a full PubMed record.
- `sort` (string; optional; default relevance; one of relevance, pub_date, author, journal): Result order.
- `date_from` (string | null; optional; pattern ^\d{4}(/\d{2}(/\d{2})?)?$): Earliest publication date: YYYY, YYYY/MM or YYYY/MM/DD.
- `date_to` (string | null; optional; pattern ^\d{4}(/\d{2}(/\d{2})?)?$): Latest publication date, same formats as `date_from`.

Base URL: `https://api.scrapercompany.com`. Auth: send your key in the `x-api-key` header. A failed request (4xx or 5xx) or an empty result costs 0 credits.

- [Full API reference as one Markdown file](https://scrapercompany.com/llms-full.txt)
- [OpenAPI spec](https://scrapercompany.com/openapi.json)
