Google Patents scraper and API skills
Site: patents.google.com
Google Patents has no REST search API — the public data is published as BigQuery datasets, which is a different tool for a different job. These Google Patents scraper skills query the live search surface through the Browserless MCP and resolve publication PDFs.
Working with Google Patents
- The search endpoint requires double URL-encoding; single-encoding silently returns the global landing page instead of results.
- Publication-number URLs are case-sensitive, and the PDF URL is read from the citation_pdf_url meta tag rather than guessed.
2 Google Patents skills
- patents.google.comv1
Download a Google Patents publication PDF
Resolve and download the PDF for a Google Patents publication using its publication number, with a caller-specified output filename.
- patents.google.comv1
Google Patents Search
Search Google Patents (~140M+ worldwide publications) by free-text query, assignee, inventor, country, status, language, or date range — returns publication number, title, priority/filing/grant/publication dates, inventor, assignee, snippet, PDF URL, and canonical landing-page URL for each match.
Google Patents scraping questions
- Does Google Patents have an API?
- Not for search. Google publishes patent data as BigQuery datasets, but there is no REST search endpoint behind patents.google.com.
- Is scraping Google Patents legal?
- Reading data that Google Patents publishes publicly is generally treated as lawful in the US — hiQ v. LinkedIn found that scraping a public site does not by itself violate the Computer Fraud and Abuse Act. That is not the whole question: Google Patents's terms of service, its robots.txt, the rate at which you request pages, and any personal or copyrighted data you collect all carry their own obligations, and the rules differ outside the US. These skills read public pages at human-scale rates and never bypass a login. Treat this as background, not legal advice.
- What data can I extract from Google Patents?
- Publication number, title, assignee, filing and publication dates, total result counts, and the resolved PDF URL.