Google Patents Search Pro

Search and fetch full patent records across USPTO, EPO, WIPO, JPO, CN, KR and 100+ patent offices. Two modes: free-text search with hard filters for country, status, patent type and priority-date range, and full records for specific patent IDs with complete abstracts, claim texts, backward and forward citations with totals, and CPC codes. Clean JSON rows, no API key needed.

Why this one

Most patent actors in the store advertise more modes than they deliver, return fewer results than the count you asked for, or hide their pricing. Google Patents Search Pro paginates properly and stops exactly at your max count, applies every filter server-side so non-matching rows are never fetched or billed, and publishes its pricing upfront: $2.00 per 1,000 patent records plus $0.005 per run.

The store's current patent-discovery leader returns partial pages past result 100 and carries zero verified reviews. This actor was built to fix exactly those two things: correct pagination to any count up to 1,000, and rows you can verify in the dataset before you pay beyond the run start.

Comparison

	Google Patents Search Pro	Current store leader
Monthly users	New actor	~33
Verified reviews	Launching with full public README and tables	0
Price per 1,000 records	$2.00	$2.00
Run start fee	$0.005	hidden
Max results per run	1,000, paginated exactly	partial pages past 100
Filters applied server-side	Yes, non-matching rows never billed	partial
Pricing published	Yes	partial

Input

Field	Type	Purpose
mode	string	search (default) to discover patents by query, or details to fetch full records for specific patent IDs.
searchQuery	string	Free-text query. Supports quoted phrases, boolean operators and wildcards, e.g. "solid state battery" electrolyte.
patentIds	array[string]	Publication numbers for details mode, e.g. US10517421B2, EP3612345A1.
countryCodes	array[string]	Restrict to offices: US, EP, WO, JP, CN, KR, DE and more. Empty means all.
status	string	GRANT for granted only, APPLICATION for pending only, empty for any.
patentType	string	PATENT for utility, DESIGN for design, empty for any.
dateFrom / dateTo	string	Priority-date bounds in YYYY-MM-DD. Empty means unbounded.
sort	string	Empty for relevance, new for newest first.
depth	string	basic (default, fast search rows) or deep (adds claims, citations, CPC per patent).
maxItems	integer	Hard cap on dataset items. Default 10, maximum 1,000.

Output

Each dataset item is one JSON row. In basic mode:

Field	Description
type	patent
patentId	Canonical publication number, e.g. US10517421B2.
title	Published title.
abstract	Abstract snippet.
priorityDate / filingDate / grantDate / publicationDate	ISO dates.
assignee	Primary assignee.
inventors	Inventor names as an array.
countryCode / kindCode	Office and publication kind codes parsed from the ID.
language	Publication language.
patentUrl	Link to the patent on Google Patents.
pdfUrl	Direct link to the published PDF.

In deep mode each row is a patent_detail and adds:

Field	Description
abstract	Complete abstract.
claimsCount / claims	Number and full text of claims.
cpc	CPC classification codes.
citationsBackward / citationsBackwardTotal	Prior-art citations and count.
citationsForward / citationsForwardTotal	Later patents citing this one, and count.

Pricing

Pay per event, published upfront:

• Actor start: $0.005 per run
• Patent record: $0.002 per dataset item, i.e. $2.00 per 1,000 patent records

A 10-record test costs about $0.025 all-in. Non-matching rows are never fetched, so filtered runs cost only for matching records.

Use cases

• Prior-art search: screen a technology area before filing, with hard date and office filters.
• IP landscaping: map who is patenting what across jurisdictions for an assignee or CPC area.
• Competitive intelligence: watch a competitor's newest filings with sort: new on a schedule.
• M&A and freedom-to-operate: pull full claims and citation networks for candidate portfolios.
• AI agents and RAG: clean structured patent rows for MCP clients and RAG pipelines.

FAQ

Which offices are covered? Any office indexed on Google Patents: USPTO, EPO, WIPO, JPO, KIPO, CNIPA, DPMA and 100+ more, selected by countryCodes.

Does it need an API key or account? No. Provide an input, run it, read the dataset.

What is the difference between basic and deep? Basic is one fast row per patent from search results. Deep fetches each patent's complete record: full abstract, claim texts, backward and forward citations with totals, and CPC codes. Deep is per-patent work, so it costs the same per record but takes longer.

Can I fetch details for patents I found elsewhere? Yes. Switch to details mode with a list of publication numbers.

How large can a run be? Up to 1,000 records per run. Pagination stops exactly at maxItems, so you are never billed for rows you did not ask for.

What date does dateFrom/dateTo filter on? Priority date, the earliest claimed filing date of the patent family.