Crossref Recent Publications by Query

For each query, pull recently-published works from Crossref's public DOI registry. One row per work, with a server-side date filter so even narrow daily windows are fast.

What you get

{
  "doi":           "10.1145/3674892.3675012",
  "title":         "Differentially Private In-Context Learning",
  "type":          "proceedings-article",
  "publisher":     "Association for Computing Machinery",
  "container":     "Proceedings of the 31st ACM SIGKDD Conference",
  "shortContainer":"KDD '26",
  "issn":          ["2153-1234"],
  "authors":       ["Alice Doe", "Bob Smith"],
  "authorCount":   2,
  "published":     "2026-05-30",
  "created":       "2026-06-01",
  "abstract":      "We study DP-SGD applied to ICL...",
  "subject":       ["Computer Science"],
  "referenceCount":     42,
  "isReferencedByCount":  0,
  "licenseUrl":    "https://www.acm.org/publications/policies/...",
  "funders":       ["NSF", "DARPA"],
  "url":           "https://doi.org/10.1145/3674892.3675012",
  "query":         "differential privacy LLM"
}

Why this matters

PubMed covers biomedical only. arXiv covers preprints in physics, CS, math, q-bio only. Crossref covers everything else — humanities, social sciences, business, law, ecology, materials science — wherever a DOI is registered. A single query gets you the new-works feed across all of academia.

How to use

Input. A list of search queries plus a date window in days. The date filter runs server-side (no over-fetch), so a 1-day window costs only as much compute as a 1-day window.

{
  "queries": ["large language models", "diffusion models"],
  "sinceDays": 7,
  "dateField": "published",
  "maxResultsPerQuery": 1000,
  "typesFilter": []
}

dateField.

• published (default) — filter on the work's print or online publication date. Best for "what came out this week."
• created — filter on when Crossref first indexed the work. Useful when a publisher backdates the print date but the DOI is fresh.

typesFilter. Restrict to specific Crossref work types. Common values: journal-article, book-chapter, proceedings-article, posted-content, dataset. Empty = no filter.

Pricing

PAY_PER_EVENT · $0.001 per paper_extracted · 1 000 papers = $1.

Buyer

• Academic research offices (literature scans).
• Bibliometrics / research-impact tools.
• Patent firms (prior-art sweep).
• Journalists tracking new policy / health / climate publications.
• Anyone building a literature-review tool that needs a working daily new-works feed.

Source

https://api.crossref.org/works — Crossref's free public REST API. The actor pages with the cursor=* deep-pagination protocol so even 100 000-result windows finish without state. UA includes a polite mailto: to land on the prioritized API pool.