Google Patents Scraper

Scrapes Google Patents for any patent query. Extracts title, patent number, dates, inventors, assignees, patent office, status, snippet, abstract, claims, citation counts, PDF URL, CPC/IPC codes, and more.

What this Actor collects

The Actor converts Google Patents results into one clean JSON record per patent or application, including publication identifiers and dates, inventors, assignees, office and status, abstract and claims, citations, classifications, document links, and search provenance when available.

• Uses the input limits and filters below to control the crawl.
• Stores source-backed fields defined by the 54-field dataset schema.
• Omits optional fields when the source does not expose a value instead of writing nulls or fabricated placeholders.

Use cases

• Patent-landscape and prior-art research
• Inventor, assignee, and citation analysis
• Classification and document-link enrichment

Input

Provide input in JSON. Fields marked required must be supplied. The Default / example column shows a schema default when one exists; otherwise it shows a documented prefill or fixture value.

Field	Type	Required	Default / example	Description
query	string	Yes	"artificial intelligence"	The search query to use on Google Patents.
maxItems	integer	No	50	Maximum number of patent results to scrape.
country	string	No	—	Filter patents by country code (e.g. US, EP, WO, CN, JP, KR, DE, GB).
dateFrom	string	No	—	Filter patents filed on or after this date (YYYYMMDD format, e.g. 20200101).
dateTo	string	No	—	Filter patents filed on or before this date (YYYYMMDD format, e.g. 20231231).
proxyConfiguration	object	No	{"useApifyProxy":true}	Proxy configuration for the scraper.

Example input

{
  "query": "machine learning",
  "maxItems": 50,
  "country": "us",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": [
      "GOOGLE_SERP"
    ]
  }
}

Output

The default dataset contains one item per patent or patent-application result. The following are the most useful fields; claims, citations, classifications, family data, and document links depend on the indexed patent record.

Field	Type	Description
position	integer	Position
title	string	Title
patentNumber	string	Patent Number
inventor	string	Inventor
assignee	string	Assignee
publicationDate	string	Publication Date
patentStatus	string	Patent Status
searchQuery	string	Search Query
scrapedAt	string	Scraped At
type	string	Type
description	string	Description
snippet	string	Snippet
url	string	URL
link	string	Link
query	string	Query
resultType	string	Result Type

Example dataset item

This compact example is taken from local Actor storage. Long text and nested collections are shortened for documentation only.

{
  "position": 1,
  "title": "Machine learning engine for fraud detection during cross-location online …",
  "patentNumber": "US10977654B2",
  "inventor": "Dinesh Kumar",
  "assignee": "Paypal, Inc.",
  "publicationDate": "2021-04-13",
  "patentStatus": "granted",
  "searchQuery": "machine learning",
  "scrapedAt": "2026-07-23T11:50:15.827Z",
  "type": "patent",
  "description": "A machine learning engine for fraud detection related to cross-location online transaction processing may be trained using artificial intelligence techniques and used according to…",
  "snippet": "A machine learning engine for fraud detection related to cross-location online transaction processing may be trained using artificial intelligence techniques and used according to…"
}

Related Actors

• Google About This Result Scraper
• Google AI Overview Scraper
• Google Autocomplete Scraper