📜 USPTO Patent Grants Tracker — IP, Legal & Competitive Intelligence

The institutional IP-monitoring desk on tap. Track newly granted US patents from the USPTO with title, assignee, inventors, CPC classifications, claims count, citation count, full-text search, AI / software flags, PDF and Google Patents links. Built for IP law firms, patent attorneys, M&A diligence teams, competitive intelligence analysts, R&D directors, tech-trend researchers, and AI / biotech / hardware funds.

This is the actor a single LexisNexis PatentSight seat costs $30k/yr to access. Pay-per-result on Apify — no seat, no contract, no minimum commitment.

---

Why a dedicated patent-grants actor (and not just a generic web scraper)

The USPTO publishes ~6,500 patent grants every Tuesday. The raw firehose is too noisy for any human workflow. A generic web scraper fails patent monitoring three ways:

1. No semantic classification. USPTO grants carry CPC (Cooperative Patent Classification) codes — the global classification standard maintained jointly by USPTO and EPO. A neural-net patent (G06N3/08), a CRISPR patent (C12N15/11), and a lithium-ion battery patent (H01M10/0525) live in completely different worlds. A generic scraper hands you all 6,500 grants in a single bucket; we hand you a CPC-typed, AI-flagged, software-flagged feed where every record is already routed to the right downstream consumer.
2. No assignee-level filtering. When the General Counsel of a hyperscaler asks "what did Google patent this week?" you can't grep raw HTML — you need normalized assignee names, fuzzy matching, country-of-record. Our assignee_name filter handles partial matches case-insensitively ("Apple" matches "Apple Inc.", "Apple Computer, Inc.", "Apple Industries").
3. No claims / citations count. The two single most-valuable patent quality signals after the assignee are (a) how many independent claims the patent asserts and (b) how many subsequent patents cite it (the forward-citation count is the gold-standard impact metric in the patent-econometrics literature). A generic scraper gives you the title; we give you a record that already includes both.

---

What you get

A unified patent grant record stream, one row per granted USPTO patent:

{
  "patent_number": "US12438891B1",
  "patent_title": "Anomaly detection based on ensemble machine learning model",
  "grant_date": "2025-10-07",
  "priority_date": "2015-08-31",
  "filing_date": "2022-02-18",
  "inventors": ["Sudhakar Muddu", "Christos Tryfonas", "Joseph Auguste Zadeh"],
  "assignee_name": "Cisco Technology Inc",
  "assignee_country": "US",
  "cpc_classes": ["G06N20/00", "G06N3/08", "H04L63/1416", "G06F21/55"],
  "abstract_excerpt": "A security platform employs a variety techniques and mechanisms to detect security related anomalies and threats in a computer network environment. The security platform is big data driven and employs machine learning to perform security analytics...",
  "claims_count": 22,
  "citations_count": 3,
  "is_ai_related": true,
  "is_software": true,
  "pdf_url": "https://patentimages.storage.googleapis.com/pdfs/US12438891B1.pdf",
  "google_patents_url": "https://patents.google.com/patent/US12438891B1/en",
  "data_source": "google_patents"
}

That single record gives you everything to:

• Track a competitor's weekly grant velocity (assignee_name + date_range: week)
• Build a quarterly AI-patent dashboard (ai_only: true + date_range: quarter — equivalent of running date_range: month three times)
• Surface freedom-to-operate risks for an R&D team's new feature (search by keywords of the contested feature)
• Build an M&A diligence patent-portfolio inventory (assignee_name of the target, date_range: year)
• Power a litigation-watch dashboard — newly granted patents in a CPC class your client is being sued under
• Run patent-econometrics studies (claims_count, citations_count, cpc_classes)
• Feed a hedge fund's R&D-momentum signal (which companies are filing the most software / AI patents and how is the cadence trending)

---

Input filters

Input	Type	Description
keywords	string list	Free-text search across title + abstract. Multi-word entries treated as exact phrases. OR'd together.
assignee_name	string	Partial match on the company / institution holding the grant. Case-insensitive.
date_range	enum	today / week / month / year / all — grant-date window.
cpc_class	string	CPC code at any level — G (section), G06N (class), G06N3/08 (subgroup).
ai_only	boolean	Filter to patents flagged AI / ML via keyword heuristic + G06N* CPC.
limit	integer	Max records to emit (1-5000).
enrich_details	boolean	If true, fetch each patent's detail page to populate cpc_classes, claims_count, citations_count, full abstract. Default true.

---

CPC class quick-reference (top sections for tech filers)

CPC	Coverage	Typical assignees
G06N	Computing arrangements based on specific computational models (AI / ML / neural nets)	Google, Meta, NVIDIA, OpenAI, Anthropic
G06F	Electric digital data processing (general software)	Microsoft, Apple, Oracle, IBM
G06Q	Data processing for business/finance/management (fintech, e-commerce, ad-tech)	Visa, Mastercard, Amazon, Salesforce
H04L	Transmission of digital information (networking, comms, blockchain)	Cisco, Ericsson, Qualcomm
H04W	Wireless networks (5G, WiFi)	Qualcomm, Samsung, Huawei, Apple
A61K	Pharmaceutical compositions	Pfizer, Merck, Moderna, Genentech
C12N	Microorganisms / biotech (CRISPR, gene editing)	Broad Institute, MIT, biotech startups
H01M	Batteries / fuel cells / capacitors	Tesla, Panasonic, CATL, LG Chem
H01L	Semiconductor devices	TSMC, Intel, Samsung, SK Hynix
B60W	Vehicle systems (ADAS, autonomous driving)	Tesla, Waymo, GM, Mobileye

Combine cpc_class: G06N + date_range: week for a weekly AI grants feed. Combine cpc_class: C12N + date_range: month for monthly biotech grants. Combine cpc_class: H01M + assignee_name: Tesla for Tesla's latest battery patents.

---

Use cases for the buyer pool

IP law firms / patent prosecutors

Daily fresh pull filtered by the client's CPC classes. Monitor competing assignees' filing cadence. Identify newly granted patents that may impact ongoing prosecution strategy. Replace your manual weekly USPTO bulk-download workflow with a JSON feed routed straight to your docketing system.

Competitive intelligence / corporate IP departments

Track every new grant by your top 20 competitors. Set up a daily scheduled task in Apify with assignee_name: <competitor> + date_range: week. The Wednesday-morning report lands in your Slack via the CI team's Apify-webhook integration before the legal team's standup.

M&A diligence teams

Acquisition target's full patent portfolio in 30 seconds: assignee_name: <target> + date_range: year + limit: 5000. Cross-reference with the target's claimed innovation pipeline. Identify whether their stated AI / biotech / hardware moat is real or marketing fluff. Add the resulting JSON to your data-room workbook.

AI / biotech / hardware venture capital

Pre-vet a portfolio startup's claimed IP before lead-check signature. Confirm their team's named inventor history. Surface adjacent assignees in the same CPC subgroup to map the competitive landscape. Track AI-grant velocity industry-wide via ai_only: true to time fund deployments to AI-grant cycle peaks.

Tech-trend researchers / industry analysts

Quarterly AI-grants benchmark: pull ai_only: true + date_range: quarter for the trailing four quarters, plot the curve. The shape of that curve is one of the half-dozen most-watched leading indicators of where corporate R&D dollars are flowing. Goldman, Morgan Stanley, Gartner, IDC, and CB Insights all publish reports based on substantially the same data — you can do it yourself for 1% of the cost.

Litigation watch / patent troll early-warning

Patent assertion entity (PAE) filings cluster in specific CPC subgroups. Monitor those subgroups weekly for newly granted patents being assigned to known PAE shell entities. Pre-position your defensive strategy 6-18 months before an infringement letter lands.

R&D / freedom-to-operate (FTO)

Before shipping a new feature, run keywords: ["<feature description>"] + date_range: year to surface fresh patents in your space. Catch FTO risks before they catch you. Cross-reference each hit's CPC classes with your existing patent portfolio's CPC classes to map gaps.

Pharma / biotech compound watch

Drug development teams use cpc_class: A61K + keywords: ["<compound or mechanism>"] to watch for newly granted compositions, formulations, and method-of-use patents in their therapeutic area. Critical for both ANDA Paragraph IV strategy and originator IP exclusivity planning.

---

Pricing — pay per patent record

Event	Cost
Actor start	$0.01
Per patent record returned	$0.10

A 5-record smoke test costs $0.51 total. A 100-record daily competitive-intel sweep costs $10.01/day or roughly $3,600/year — about 12% of a single LexisNexis PatentSight seat ($30k/yr), with full programmatic access and zero seat licensing.

Patent records are priced at $0.10 each (5x our regional stock screeners at $0.02 and 2x our SEC Form 4 at $0.10) because each patent is a discrete, high-value, defensible asset. A single granted patent in a hot CPC subgroup can drive a 6-figure litigation outcome, a 7-figure licensing deal, or a 9-figure M&A repricing. We anchor price to event value, not byte count.

---

Data sources & methodology

Primary: Google Patents search index. Google Patents indexes every newly granted USPTO patent within ~24 hours of publication via the USPTO Bulk Data Service feed. The Google Patents JSON XHR endpoint returns structured records with patent number, title, abstract snippet, primary inventor, primary assignee, grant date, priority date, filing date, and PDF link. We hit the search endpoint with your filters and paginate as needed.

Enrichment: per-patent detail pages. When enrich_details: true (default), the actor fetches each patent's Google Patents detail page and extracts the full CPC class list, the claims count (via id="CLM-NNNNN" anchor count on the rendered claims block), the forward-citation count (Cited by marker), and the full abstract. Detail enrichment doubles run time but produces a far richer record. For high-volume scans where only the search-result fields are needed, set enrich_details: false.

Classification heuristics:

• is_ai_related fires if (a) the title or abstract contains any of ~30 AI / ML keywords (machine learning, neural network, transformer, large language model, generative AI, etc.) or (b) any CPC class starts with G06N (computing arrangements based on specific computational models — the canonical AI/ML CPC class).
• is_software fires if the title / abstract contains software-specific terms (computer-implemented, API, algorithm, etc.) or any CPC class is G06F, G06Q, or H04L.

Underlying data validity. Every record carries google_patents_url and pdf_url so you can audit the source patent in two clicks. Patent numbers are canonical USPTO publication numbers (e.g. US12438891B1) — paste them straight into PAIR (Patent Application Information Retrieval) or Public PAIR for the official wrapper.

Coverage scope. US-issued utility patents and design patents are returned. We default to country=US and status=GRANT to exclude pre-grant publications (kind code A1) and non-USPTO grants. Set date_range: all to remove the time-window and pull arbitrary historical grants back to ~1976 (the start of USPTO's full-text database).

---

How to run

Apify Console — paste your filter inputs, hit Run, download the dataset as JSON or CSV.

Apify API (Python):

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("nexgendata/uspto-patent-grants").call(run_input={
    "keywords": ["large language model", "transformer"],
    "assignee_name": "",
    "date_range": "month",
    "cpc_class": "G06N",
    "ai_only": True,
    "limit": 100,
    "enrich_details": True
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["patent_number"], "—", item["assignee_name"],
          "—", item["patent_title"][:80], "—", item["claims_count"], "claims")

Schedule it — Apify Scheduled Tasks. The USPTO publishes new grants every Tuesday morning ET, so a Tuesday 09:00 ET schedule captures the fresh weekly drop. For competitive-intel use cases, run nightly with date_range: today and pipe the diff into Slack / Notion / a Snowflake table.

---

FAQ

How fresh is the data? Google Patents typically indexes new USPTO grants within 12-24 hours of the official Tuesday publication. The actor pulls live from the Google Patents search index — no caching layer in between.

What about pending (not-yet-granted) applications? This actor returns granted patents only (status=GRANT). For pre-grant publications (kind code A1 — applications published 18 months after filing, before grant), use our uspto-patent-applications actor (coming soon). For provisional applications, those are not published by USPTO until / unless they convert to a non-provisional filing.

Can I get the full claims text? Currently claims_count is emitted but the full claims text is not — it would push payload size 10-50x and isn't reliably extractable from the search-result JSON. For the full claims, follow the google_patents_url or pdf_url in each record.

Does it cover design patents and plant patents? Utility and design patents are returned (USPTO publication numbers starting with US plus 7-8 digits). Plant patents (USPP prefix) are included if they match your search. Reissue patents (USRE prefix) are included. Provisional applications are excluded (they're not published).

Does it work for non-US jurisdictions (EP, WO, CN)? This actor restricts to US patents (country=US). For European patents use our roadmap EPO actor; for WIPO PCT applications use the WIPO Patentscope actor; for Chinese patents use CNIPA-direct via our China actor fleet.

Why CPC and not USPC? USPTO retired its legacy US Patent Classification (USPC) system in 2015 and now classifies all new grants in CPC (Cooperative Patent Classification, the joint USPTO-EPO standard). All modern grants carry CPC codes; USPC is legacy-only.

How do I find the CPC code I want? Search the CPC scheme at https://www.uspto.gov/web/patents/classification/cpc/html/cpc.html or use the Google Patents UI to find a representative patent in your area, note its CPC, and use that as your filter.

Can I monitor multiple assignees simultaneously? Run the actor once per assignee in parallel via the Apify API; or use keywords to OR-search across assignee names embedded as keywords (less precise but faster). For systematic competitive-intel workflows we recommend the parallel-call pattern with a scheduled task per competitor.

---

NexGenData vs. paid patent intelligence platforms

Tool	Annual cost	API access	USPTO coverage	AI classification
LexisNexis PatentSight	~$30,000/seat	Paid add-on	Full	Manual taxonomy
Clarivate Derwent Innovation	~$25,000/seat	Paid add-on	Full	Editorial tagging
Patsnap	~$15,000/seat	Paid add-on	Full	AI-tagged
Questel Orbit Intelligence	~$20,000/seat	Paid add-on	Full	Topic clustering
NexGenData USPTO Patent Grants	Pay-per-result	Native (REST + MCP)	Full USPTO	CPC + keyword AI flag

Run this actor at 100 records/day, 250 trading days/year, and your annual cost is approximately $3,600 — an 85-93% discount vs. the major paid alternatives, with no seat licensing, no minimum commitment, and full programmatic access via Apify's REST API or any of 1,500+ MCP / no-code / Zapier / Make integrations.

The trade-offs vs. the paid platforms:

• ✅ Same primary source data (USPTO grants via Google Patents index — Google Patents is the same data underneath several of the paid platforms' UIs).
• ✅ Same CPC classification — the global standard.
• ✅ Full programmatic access — no rate limits, no per-seat licensing.
• ❌ No proprietary "patent quality score" — but we emit claims_count + citations_count, the two raw signals the proprietary scores wrap.
• ❌ No built-in family graph (related applications across jurisdictions) — for global patent families, layer this actor with our roadmap EPO + WIPO actors.
• ❌ No litigation tagging / standards-essential-patent (SEP) flagging — we emit the raw patent record; pair with a litigation database for the overlay.

For 80% of patent-monitoring workflows — competitive cadence tracking, FTO checks, M&A diligence, R&D trend research — this actor is functionally equivalent to the paid platforms at 7-15% of the cost.

---

Built by NexGenData

We build data infrastructure for global financial, legal, and IP intelligence desks. 220+ actors live on Apify covering global equities, IPOs, SEC filings, institutional holdings, M&A, patents, B2B intelligence, and MCP servers for LLM integration. Patent data complements our broader IP-and-legal stack — combine this actor with our SEC filings trackers, IPO Tracker, and institutional-holdings actors to build a fully connected IP-through-public-markets data pipeline.

---

Explore the NexGenData IP & Capital Markets Suite

Patent data sits at the intersection of R&D, capital markets, and corporate strategy. Pair the USPTO Patent Grants Tracker with NexGenData's IPO, institutional-holdings, activist, and startup-funding trackers to build a fully connected R&D-through-public-markets pipeline:

Institutional capital intelligence

• 🔗 SEC Form 13F Holdings Tracker — Quarterly hedge-fund and mutual-fund positions. Cross-reference patent-grant velocity with institutional accumulation.
• 🔗 SEC Schedule 13D/G Activist Tracker — 5%+ activist positions. Detect activists targeting patent-rich tech / biotech companies.

Capital-markets lifecycle

• 🔗 IPO Tracker — Recent and upcoming IPOs. Combine with this actor to inventory the S-1 issuer's full patent portfolio at filing time.
• 🔗 Startup Funding Tracker — Pre-IPO funding rounds. Tie each round to the company's named inventor history.

Developer / tech-trend leading indicators

• 🔗 GitHub Trending Repos — Daily GitHub-velocity ranking. Patent grants are a lagging IP signal; GitHub trends are a leading R&D signal. Combine the two to map the full innovation pipeline.

Multi-asset / specialty

• 🔗 Finance MCP Server — Unified Model Context Protocol server exposing the full NexGenData IP and capital-markets fleet to Claude, Cursor, and any MCP-aware LLM client.

Browse the full NexGenData catalog

→ See all 220+ NexGenData actors at apify.com/nexgendata — IP intelligence, equity screeners across 25+ markets, IPO + pre-IPO trackers, SEC filings, institutional holdings, B2B intelligence, MCP servers, and more.

---

---

Quick Comparison vs. Paid Incumbents

How this actor stacks up against the platforms patent and competitive-intelligence teams pay six figures for:

Capability	NexGenData USPTO Patent Grants Tracker	Bloomberg Terminal	PitchBook	Crunchbase Pro	Refinitiv
Annual cost	Pay-per-result (~$3,600 typical)	$24k+/seat	$30–80k/seat	$24k+/seat	$22k+/seat
Dedicated USPTO patent grants feed	Yes — native, full history	Indirect / news	Indirect / news	Indirect / news	Indirect / news
CPC classification + AI flag emitted per row	Yes	No	No	No	No
Programmatic API (REST + MCP)	Yes — included	Paid add-on	Paid add-on	Paid add-on	Paid add-on
Seat license / annual minimum	None	$24k+	$30k+	$7k+	$22k+
Self-service, no sales call	Yes	No	No	Partial	No

For 80% of patent-monitoring workflows — competitive cadence tracking, FTO checks, M&A diligence, R&D trend research — this actor delivers parity coverage at 5–15% of incumbent cost.

---

Quick Comparison vs. Paid Incumbents

How this actor stacks up against the platforms patent and competitive-intelligence teams pay six figures for:

Capability	NexGenData USPTO Patent Grants Tracker	Bloomberg Terminal	PitchBook	Crunchbase Pro	Refinitiv
Annual cost	Pay-per-result (~$3,600 typical)	$24k+/seat	$30–80k/seat	$24k+/seat	$22k+/seat
Dedicated USPTO patent grants feed	Yes — native, full history	Indirect / news	Indirect / news	Indirect / news	Indirect / news
CPC classification + AI flag emitted per row	Yes	No	No	No	No
Programmatic API (REST + MCP)	Yes — included	Paid add-on	Paid add-on	Paid add-on	Paid add-on
Seat license / annual minimum	None	$24k+	$30k+	$7k+	$22k+
Self-service, no sales call	Yes	No	No	Partial	No

For 80% of patent-monitoring workflows — competitive cadence tracking, FTO checks, M&A diligence, R&D trend research — this actor delivers parity coverage at 5–15% of incumbent cost.

🚀 Apify affiliate program — Get free credits & save 30%

Sign up to Apify via our referral link and you'll get:

• Free starter credits to test this actor and the rest of our 220+ actor fleet
• 30% off platform fees for the life of your account

→ Sign up here with NexGenData's affiliate link — same Apify, same actors, just cheaper for you and a small kickback that helps us keep building.

Built and maintained by NexGenData — institutional-grade IP, equity, IPO, SEC filings, and B2B data infrastructure on Apify.