USPTO Patent Search Scraper

Search and extract US patent data from the USPTO Public Patent Search (PPUBS) — the official full-text patent search system covering 10+ million US patents and published applications. No API key required.

What you can scrape

• Patents — titles, publication numbers, filing dates, publication dates, assignees, inventors, CPC/IPC classification codes
• Patent applications — published pre-grant applications (US-PGPUB)
• Granted patents — issued US patents (USPAT)
• Full patent details — abstract, claims page count, examiner info, priority dates

Use cases

• Patent research and prior art searches
• Competitor patent portfolio analysis
• Technology trend analysis by keyword or date range
• Inventor or company patent activity tracking
• CPC/IPC classification-based research
• Academic research on patent innovation

Input

Field	Type	Description
searchQuery	string	Keywords to search in patent titles and abstracts
assigneeName	string	Filter by patent owner (company or individual)
inventorName	string	Filter by inventor name
dateFrom	string	Start date filter (YYYY-MM-DD)
dateTo	string	End date filter (YYYY-MM-DD)
patentType	select	Any, Utility, Design, Plant, or Reissue
sortBy	select	Newest first, Oldest first, or Best match
fetchDetails	boolean	Fetch full details including abstract (slower)
maxItems	integer	Max records to return (1–1000, default 25)

Example inputs

Search AI patents:

{
  "searchQuery": "artificial intelligence",
  "maxItems": 50
}

Search Google's recent patents:

{
  "searchQuery": "machine learning",
  "assigneeName": "Google",
  "dateFrom": "2022-01-01",
  "maxItems": 100
}

Find patents by inventor with full details:

{
  "inventorName": "Elon Musk",
  "fetchDetails": true,
  "maxItems": 25
}

Search design patents:

{
  "searchQuery": "electric vehicle",
  "patentType": "design",
  "maxItems": 50
}

Date range search for CRISPR patents:

{
  "searchQuery": "CRISPR gene editing",
  "dateFrom": "2020-01-01",
  "dateTo": "2024-12-31",
  "maxItems": 200
}

Output

Patent record

{
  "guid": "US-20260147587-A1",
  "publicationNumber": "20260147587",
  "documentId": "US 20260147587 A1",
  "title": "SYSTEM AND METHOD FOR AUTOMATED PLANNING ASSISTANT",
  "kind": "A1",
  "documentType": "US-PGPUB",
  "publicationDate": "2026-05-28",
  "filingDate": "2024-09-24",
  "assignees": ["JPMorgan Chase Bank, N.A."],
  "inventors": ["BORRAJO; Daniel et al."],
  "applicationNumber": "18/894868",
  "cpcCodes": ["G06F9/451", "H04L51/02"],
  "ipcCodes": ["G06F9/451", "H04L51/02"],
  "abstract": "Methods and systems for automated planning using AI...",
  "pageCount": 22,
  "claimsPages": 2,
  "sourceUrl": "https://ppubs.uspto.gov/pubwebapp/external.html?q=(20260147587).pn.&db=US-PGPUB",
  "scrapedAt": "2026-05-30T10:30:00+00:00",
  "recordType": "patent"
}

FAQ

Do I need an API key? No. The USPTO PPUBS API is publicly accessible without authentication or registration.

What databases are searched? All three USPTO databases: US-PGPUB (published applications), USPAT (granted patents), and USOCR (legacy patents).

How is the search performed? Searches use USPTO Boolean Retrieval Syntax (BRS) — the same syntax used in the PPUBS web interface. Keywords are searched in both patent titles and abstracts.

What does fetchDetails do? When enabled, the scraper fetches the full patent record for each result, adding abstract text, claims details, and more. This is slower but returns more data per patent.

Why are some fields missing? Only fields with actual data are included. Empty, null, or blank fields are never emitted.

What is the difference between US-PGPUB and USPAT? US-PGPUB contains published patent applications (pre-grant). USPAT contains issued/granted patents. A patent application may first appear in US-PGPUB and later in USPAT after grant.

How many patents are searchable? The USPTO PPUBS database contains over 10 million US patents and patent applications.

What sort options are available?

• Newest first (default): most recently published patents
• Oldest first: earliest published patents
• Best match: relevance-ranked by search score