Google Patents Scraper (Search, Abstracts, CPC)

What is Google Patents Scraper?

Google Patents Scraper searches Google Patents by keyword, assignee, inventor, country, status and publication date. It returns one row per patent: publication number, title, snippet, assignee, inventors, priority, filing and publication dates, country, PDF link and the Google Patents page. Optionally it opens each patent page for the abstract, claims, CPC classification codes and legal status. You can also skip the search and look up specific patent numbers.

Use it for prior-art searches, competitor tracking by assignee, technology scans by CPC topic and patent lists for research. It reads the public Google Patents search endpoint and patent pages. No Google login is used.

What data does Google Patents Scraper extract?

Field	Type	Meaning
publicationNumber	string	Patent or application number, e.g. US9123456B2
position	number	Rank within the search (1 = first)
page	number or null	Search page the row came from (100 per page)
query	string or null	The keywords used
title	string	Patent title
snippet	string or null	Excerpt from the search result
assignee	string or null	Current assignee as Google shows it
inventors	string or null	Inventors, separated by semicolons
priorityDate, filingDate, grantDate, publicationDate	string or null	ISO dates (YYYY-MM-DD)
country	string	Office code from the number, e.g. US, CN, EP
language	string or null	Language of the record
url	string	Google Patents page for the patent
pdfUrl	string or null	Link to the patent PDF
abstract	string or null	Abstract (with includeDetails or patent numbers)
claimCount, claims	number or null, string or null	Number of claims and their text, up to 20,000 characters
cpcCodes	string or null	Leaf CPC codes, separated by semicolons
legalStatus	string or null	Status Google shows, e.g. Active, Expired - Fee Related
scrapedAt	string	ISO timestamp of the scrape

How to use Google Patents Scraper

1. Enter keywords in "Search keywords", e.g. solid-state battery.
2. Optional: add an assignee, an inventor, a country code, a status or a publication date range.
3. Set "Max patents" (default 50, up to 500).
4. Turn on "Include full record" if you need the abstract, claims, CPC codes and legal status.
5. Or leave the keywords empty and list patent numbers under "Patent numbers" to look up those patents only.
6. Start the run and download the dataset, or read it through the API.

Example input

{
    "query": "solid-state battery",
    "assignee": "Tesla",
    "status": "GRANT",
    "maxResults": 20,
    "includeDetails": true
}

Example output

Real row from a patent-number lookup on 23 Oct 2026 (abstract and claims shortened here):

{
    "publicationNumber": "US9123456B2",
    "position": 1,
    "title": "High frequency cable, high frequency coil and method for manufacturing high frequency cable",
    "assignee": "Fujikura Ltd",
    "inventors": "Takashi Shinmoto; Shotaro Yoshida; Shinji Katayama; Taikou TODA; Takamasa Kato; Masanori DAIBO; Akio Kawakami",
    "priorityDate": "2010-03-23",
    "filingDate": "2012-09-21",
    "grantDate": null,
    "publicationDate": "2015-09-01",
    "country": "US",
    "language": "en",
    "url": "https://patents.google.com/patent/US9123456B2/en",
    "pdfUrl": "https://patentimages.storage.googleapis.com/6c/70/d5/8682e85a6a4c05/US9123456.pdf",
    "abstract": "A high frequency cable includes: a central conductor made from aluminum or an aluminum alloy; a covering layer made from copper covering the central conductor ...",
    "claimCount": 5,
    "claims": "1. A high frequency cable, comprising: a central conductor made from aluminum or an aluminum alloy; ...",
    "cpcCodes": "H01B13/00; H01B13/0006; H01B13/06; B21C3/00; ...; H01B7/00; H01B7/30; H01F27/00; ...",
    "legalStatus": "Expired - Fee Related",
    "scrapedAt": "2026-10-23T10:18:34.674Z"
}

Input parameters

Parameter	Default	Description
query	empty (demo solid-state battery when no patent numbers are given)	Keywords
patentNumbers	empty	Patent numbers to look up, one per line
assignee	empty	Company or person that owns the patent
inventor	empty	Inventor name
country	empty	Office code: US, CN, EP, WO, JP, KR
status	any	any, GRANT or APPLICATION
publishedAfter, publishedBefore	empty	Publication date range, YYYY-MM-DD
maxResults	20	Patents per run, 1 to 500
includeDetails	false	Open each patent page for abstract, claims, CPC and status

Pricing

Pay per patent: $2.00 per 1,000 patents on the free plan ($0.002 per patent). Volume tiers lower the price on paid Apify plans.

Measured on 23 Oct 2026 at 256 MB: one run for 150 patents (keyword battery, search only) finished in 18 seconds. Platform usage is shown in the Apify run details; the figure depends on the run time, so check the run page for your own input.

Google Patents Scraper vs. alternatives

Google Patents Scraper returns what the public Google Patents search shows: its ranking, its assignee and inventor names as Google writes them, and the CPC codes from the patent page. It does not replace an official patent office bulk download or a commercial patent database with legal-status history. Use this Actor when you need a quick, filtered list from Google Patents with the full record on demand.

Using Google Patents Scraper with AI agents / MCP

An agent sends query (or patentNumbers) plus the filters and reads the dataset rows. Keep maxResults small for broad queries, because each patent is one charged row and the caller's maxTotalChargeUsd caps the total.

FAQ

Is this legal? The Actor reads public Google Patents pages and the public search endpoint that those pages use. It does not log in. Check Google's terms for your own use, and respect the rights of the patent holders for any reuse of the text or PDFs.

Why is grantDate empty for some patents? Google shows a grant date only when the record has one. Older patents and some applications have none. The publication and priority dates are still filled.

Why is the abstract empty for some patents? In one test (EP3586387B1, 19 of 20 records had an abstract), the page had no abstract in the markup the Actor reads. The row is still returned with the other fields.

How many patents can one run return? Up to 500 per run. Google reports the total hits for each query in its results, but the Actor stops at your maxResults.

Are there blocks or captchas? In four test runs from Apify, the search endpoint and the patent pages answered every request without a proxy. Retries are built in for rate limits (HTTP 429) and server errors.

Does it work with n8n, Make or Zapier? Those tools can start any Apify Actor through their Apify integrations. This was not tested for this Actor.

What about errors and limits? A run fails only when no patent is collected. Patent numbers that are not found are reported in the log and skipped. Each run stops when the caller's maximum charge is reached.

Related Actors

Other Actors from the same publisher are listed on its Apify profile.

Support

Open an issue on the Actor's Issues tab on Apify. Include the run ID, your input and the log lines that mention a problem.

Limitations

• The search returns Google's own ranking and Google's own assignee and inventor names. Names are not normalised across spellings.
• Legal status is what Google shows on the page. It is not a legal opinion, and it can lag the patent office register.
• Claims are cut at 20,000 characters per patent. Very long claim sets are truncated.
• Full records need one extra page request per patent. Runs with includeDetails take longer.