WIPO PATENTSCOPE Patent Search

Search WIPO's PATENTSCOPE database — the world's most comprehensive international patent index, with ~110 million patent documents across PCT, EPO, USPTO, JPO, KIPO, CNIPA and 50+ national offices — and get clean structured JSON.

This actor turns the WIPO PATENTSCOPE results page into a normalized dataset of publication metadata: application numbers, publication numbers, applicants (assignees), named inventors, IPC classification codes, publication dates, country of publication, and a 500-character abstract excerpt — each row carrying a source_url back to the canonical WIPO record for verification.

Built and maintained by NexGenData. Polite by design: a 2 req/s rate limit, a contact-identifying User-Agent, and a runtime robots.txt check before every job.

Why teams use this actor

• Patent landscape analysis — map all PATENTSCOPE filings in a technology area (e.g. "diffusion model training") over a year window and feed the dataset straight into pandas, Looker, or your in-house IP analytics stack.
• Freedom-to-operate (FTO) searches — pull recent publications matching a product description plus claim keywords; export to your patent counsel as a structured CSV instead of a stack of PDFs.
• Competitor IP monitoring — schedule weekly runs filtered by applicants (e.g. ["NVIDIA Corporation", "Alphabet Inc"]) to track every new PCT filing from named rivals.
• Prior art research — power R&D teams and patent examiners with a fast, structured search over the WIPO corpus, including IPC code filters and date ranges.
• M&A IP due diligence — capture every filing tied to a target company in minutes and surface the technology stack from their patent family.
• Academic / market research — quantify innovation trends by IPC class, geography, or filing entity over decades of WIPO data.

Input

Field	Type	Description
query	string (required)	Free-text query across title, abstract, claims. Supports PATENTSCOPE field operators (e.g. AI AND TI:transformer).
applicants	string[]	Optional applicant / assignee names. Combined as an OR group, then AND-ed with query.
yearStart	integer	Earliest publication year (default 2024).
yearEnd	integer	Latest publication year (default 2026).
maxResults	integer	Hard cap on patents returned (1–1000, default 25).
language	enum	Result language preference: EN, FR, DE, ES, JA, ZH, RU.

Output

One JSON object per patent, written to the actor's default dataset. Example:

{
  "application_number": "PCT/US2026/012345",
  "publication_number": "WO2026/098765",
  "publication_date": "2026-04-15",
  "title": "Method and apparatus for distributed model inference",
  "applicants": ["Acme Corp", "University of Foo"],
  "inventors": ["Smith, John", "Doe, Jane"],
  "abstract_excerpt": "A computer-implemented method for distributed inference across heterogeneous accelerators ...",
  "ipc_codes": ["G06F 21/56", "H04L 9/06"],
  "country_of_publication": "WO",
  "source_url": "https://patentscope.wipo.int/search/en/detail.jsf?docId=WO2026098765",
  "scraped_at": "2026-05-28T12:00:00Z"
}

Sample Output

(Screenshot placeholder — sample dataset rendered in the Apify console.)

Compliance

• Public bibliographic data only. Inventor names are captured as printed on the published filing; no email addresses, phone numbers, or personal mailing addresses are ever extracted.
• Polite User-Agent identifying NexGenData with a contact URL/email.
• Respects WIPO's robots.txt (checked at the start of every run).
• Rate-limited to ≤ 2 requests/second.

Keywords

WIPO patent search, PATENTSCOPE API, international patent data, patent intelligence, PCT filings, patent landscape analysis, IPC classification search, freedom to operate, prior art discovery, competitor IP monitoring.