Crossref REST API

Use in your agent

Load Crossref works for a search into pandas and Parquet. Checked against the provider's documentation on Nov 3, 2026.

Access

• No sign-up or key. Put your email address in the mailto parameter (or the User-Agent) to use the polite pool.
• Base URL https://api.crossref.org; /works takes query, filter, rows, select and cursor parameters.

Request

curl 'https://api.crossref.org/works?query.bibliographic=monetary+policy+transmission&filter=from-pub-date:2025-01-01,type:journal-article&rows=100&mailto=you@example.org'

Format

JSON with status, message-type and message; message carries total-results and items, each item a work with DOI, title (a list), container-title, published date-parts, is-referenced-by-count, publisher, and optional funder, license and abstract fields.

Load into pandas and Parquet

Needs Python 3 with requests, pandas and pyarrow.

"""Crossref works metadata -> pandas -> Parquet. No sign-up; add mailto for the polite pool."""
import requests
import pandas as pd

params = {
    "query.bibliographic": "monetary policy transmission",
    "filter": "from-pub-date:2025-01-01,type:journal-article",
    "rows": 100,
    "select": "DOI,title,container-title,published,is-referenced-by-count,publisher",
    "mailto": "you@example.org",  # replace with your own contact address
}
r = requests.get("https://api.crossref.org/works", params=params, timeout=60)
r.raise_for_status()
msg = r.json()["message"]
rows = [{
    "doi": it["DOI"],
    "title": (it.get("title") or [None])[0],
    "journal": (it.get("container-title") or [None])[0],
    "published": "-".join(str(p) for p in (it.get("published", {}).get("date-parts") or [[None]])[0] if p is not None),
    "cited_by": it.get("is-referenced-by-count"),
    "publisher": it.get("publisher"),
} for it in msg["items"]]
df = pd.DataFrame(rows)
df.to_parquet("crossref_works.parquet", index=False)
print(len(df), "works on this page of", msg["total-results"], "matches")

Run by us on 2026-11-03 (python 3.13.5 | pandas 3.1.0 | pyarrow 26.0.0 | requests 2.34.2). Result: 100 works on this page of 140686 matches

Limits

• Crossref's table: public pool 5 with concurrency 1, polite pool 10 with concurrency 3; on 2026-11-03 the headers showed 1 and 3 per 1s, so read x-rate-limit-limit and x-rate-limit-interval.
• A 429 means slow down; a manual block returns 403, which is why Crossref asks for a mailto in every request.

Quirks

• title and container-title are lists; take the first element.
• Dates come as date-parts arrays such as [[2025, 3, 14]], sometimes with only a year.
• Use cursor=* for deep paging instead of large offsets.

Terms and attribution

• Crossref says almost none of the metadata is subject to copyright and it may be used for any purpose.
• Some abstracts in the metadata may be under publishers' or authors' copyright.

Sources

• https://www.crossref.org/documentation/retrieve-metadata/rest-api/: “No sign-up is required to use the REST API, and almost none of the metadata is subject to copyright, and you may use it for any purpose.”
• https://www.crossref.org/documentation/retrieve-metadata/rest-api/access-and-authentication/: “Anyone can access the REST API, no signup or registration is required.”
• https://www.crossref.org/documentation/retrieve-metadata/rest-api/access-and-authentication/: “If you exceed the rate limits, you will receive a 429 HTTP response.”