IP Patent & Trademark MCP Server
Patent and trademark intelligence for AI agents, delivered via the Model Context Protocol. This MCP server connects Claude Desktop, Cursor, Windsurf, and any MCP-compatible client directly to three official IP databases — USPTO PatentsView for U.S. patents, EPO Open Patent Services for European patents, and EUIPO for EU trademarks — with no code required beyond a one-line config entry.
Maintenance Pulse
90/100Cost Estimate
How many results do you need?
Pricing
Pay Per Event model. You only pay for what you use.
| Event | Description | Price |
|---|---|---|
| us-patent-search | Search USPTO patent records. | $0.05 |
| eu-patent-search | Search European Patent Office records. | $0.05 |
| eu-trademark-search | Search EUIPO EU trademark records. | $0.05 |
| global-patent-search | Composite patent search across US and EU patent offices. | $0.10 |
Example: 100 events = $5.00 · 1,000 events = $50.00
Connect to your AI agent
Add this MCP server to Claude Desktop, Cursor, Windsurf, or any MCP-compatible client.
https://ryanclinton--ip-patent-trademark-mcp.apify.actor/mcp{
"mcpServers": {
"ip-patent-trademark-mcp": {
"url": "https://ryanclinton--ip-patent-trademark-mcp.apify.actor/mcp"
}
}
}Documentation
Patent and trademark intelligence for AI agents, delivered via the Model Context Protocol. This MCP server connects Claude Desktop, Cursor, Windsurf, and any MCP-compatible client directly to three official IP databases — USPTO PatentsView for U.S. patents, EPO Open Patent Services for European patents, and EUIPO for EU trademarks — with no code required beyond a one-line config entry.
The server exposes 5 tools covering U.S. patent search, European patent search, EU trademark search, and a composite cross-jurisdictional patent analysis that queries both USPTO and EPO in parallel. All three underlying data sources offer free API credentials. This server orchestrates them into a unified IP intelligence layer for your AI workflows.
What data can you access?
| Data Point | Source | Example |
|---|---|---|
| 📄 Patent title | USPTO PatentsView | "Solid-state lithium-ion battery with ceramic separator" |
| 🔬 Abstract | USPTO PatentsView | "A method for manufacturing..." |
| 👤 Inventor names | USPTO / EPO | "Chen, Wei; Nakamura, Taro" |
| 🏢 Assignee / Applicant | USPTO / EPO | "Panasonic Holdings Corporation" |
| 📅 Grant / Publication date | USPTO / EPO | "2024-03-12" |
| 🗂️ CPC classification codes | USPTO / EPO | "H01M10/0525" (battery technology) |
| 🔗 Citation count | USPTO PatentsView | 47 forward citations |
| 🌍 Publication country | EPO OPS | EP, DE, FR, GB, NL |
| 🏷️ Trademark verbal element | EUIPO | "NEXAFLOW" |
| 📦 Nice class | EUIPO | 9, 35, 42 |
| ✅ Trademark status | EUIPO | REGISTERED, FILED, OPPOSED |
| 📋 Goods and services | EUIPO | "Software for supply chain management (Class 42)" |
| 🔁 Cross-jurisdiction filing | USPTO + EPO combined | Shared CPC codes between US and EU results |
Why use the IP Patent & Trademark MCP Server?
Patent and trademark research is time-consuming, fragmented across databases with incompatible interfaces, and rarely accessible from inside the AI tools where attorneys and analysts already work. A freedom-to-operate search that touches USPTO and EPO manually can take a full day of switching interfaces, reformatting exports, and reconciling date formats.
This server removes that friction entirely. Your AI assistant can search three IP databases in a single conversation turn, deduplicate cross-border filings by CPC code, and return structured data ready for analysis.
Running this server on Apify's platform also gives you:
- Scheduling — run IP monitoring on a daily or weekly cadence to track new filings in your technology space
- API access — trigger patent and trademark searches from Python, JavaScript, or any HTTP client using the Apify API
- Standby mode — the server stays warm and responds in under a second, with no cold-start delay for interactive AI sessions
- Monitoring — receive Slack or email alerts if a run fails or returns unexpected output
- Integrations — connect results to Google Sheets, HubSpot, Zapier, or downstream actors in a pipeline
Features
- 5 MCP tools, 3 official IP databases — USPTO PatentsView, EPO Open Patent Services, and EUIPO accessed through their official APIs
- U.S. patent search with 8 filter dimensions — keyword (title full-text), inventor, assignee, date range, CPC classification prefix, patent type (utility/design/plant/reissue), sort field, and up to 500 results per query
- European patent search with raw CQL support — filter by title keywords, applicant, inventor, publication number, country code, and date range, or pass a raw CQL query for advanced searches
- EU trademark search across all 45 Nice classes — filter by verbal element, Nice class, status (REGISTERED / EXPIRED / CANCELLED / REFUSED / FILED / OPPOSED / WITHDRAWN), applicant, mark feature (WORD / FIGURATIVE / SHAPE_3D / COLOUR / SOUND / POSITION / PATTERN / MOTION / MULTIMEDIA), and date range
- Cross-jurisdictional patent analysis —
ip_global_patent_searchqueries USPTO and EPO in parallel, then identifies shared CPC 4-character code groups appearing in both result sets and common assignee/applicant names filing in both jurisdictions - Response aggregation — each tool returns structured summary metadata (total results, unique assignees, countries, status breakdowns) alongside the full result array
- Graceful credential handling — tools detect missing API credentials and return actionable setup instructions rather than errors, so onboarding is self-guided
- Pay-per-event billing — charged per tool call, not per minute; idle standby time is free
- Stateless request handling — each MCP request spawns a fresh server instance with no shared state, suitable for concurrent multi-agent use
- 120-second actor timeout — each underlying data actor is called with a 120-second timeout and 256 MB memory allocation
Use cases for IP patent and trademark search
Patent landscape analysis for technology teams
R&D teams and corporate patent attorneys need to map the competitive patent landscape before filing. ip_global_patent_search runs USPTO and EPO queries simultaneously, returns shared CPC codes to identify where a technology cluster overlaps across jurisdictions, and surfaces common assignees filing in both markets. What previously required two separate database sessions now happens in one prompt.
Freedom-to-operate research
Attorneys conducting FTO clearance studies can query ip_us_patents filtered by a specific assignee and CPC code to identify patents held by a potential blocking party, then cross-reference against ip_eu_patents for European equivalent coverage. The structured JSON output maps directly into FTO analysis workflows.
EU trademark clearance before filing
Brand teams and IP counsel checking availability before a trademark application can use ip_eu_trademarks to search by verbal element and Nice class, filter by REGISTERED status, and review the goods and services descriptions of conflicting marks. Reduces pre-filing clearance time from hours to minutes.
Competitive intelligence for investors and analysts
Investment analysts tracking a company's IP output can search ip_us_patents by assignee, sorted by patent date, to monitor the cadence of new filings. Combining this with ip_eu_patents filtered by applicant name gives a picture of whether a company is pursuing international protection on its inventions.
Inventor tracking and academic research monitoring
Technology scouts and talent acquisition teams can track a specific inventor's patent output across both U.S. and European databases, identifying their current employer (the assignee or applicant), recent filing activity, and the CPC technology domains they work in.
AI agent workflows for IP monitoring
Legal operations teams building AI agents for ongoing IP monitoring can configure this server in Claude or Cursor, then instruct the agent to run weekly searches for new filings in a given CPC class and flag any new entrants. The MCP protocol means no bespoke integration code is required.
How to connect the IP Patent & Trademark MCP Server
Claude Desktop
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"ip-patent-trademark": {
"type": "streamableHttp",
"url": "https://ip-patent-trademark-mcp.apify.actor/mcp",
"headers": {
"Authorization": "Bearer YOUR_APIFY_TOKEN"
}
}
}
}
Cursor
Add this to your Cursor MCP configuration:
{
"mcpServers": {
"ip-patent-trademark": {
"type": "streamableHttp",
"url": "https://ip-patent-trademark-mcp.apify.actor/mcp",
"headers": {
"Authorization": "Bearer YOUR_APIFY_TOKEN"
}
}
}
}
Windsurf / other MCP clients
Any client supporting the Streamable HTTP MCP transport can connect using:
- Endpoint:
https://ip-patent-trademark-mcp.apify.actor/mcp - Method: POST
- Auth:
Authorization: Bearer YOUR_APIFY_TOKEN
Your Apify token is available at console.apify.com/account/integrations.
Available MCP tools
| Tool | Source | Charge event | Description |
|---|---|---|---|
ip_us_patents | USPTO PatentsView | us-patent-search | Search U.S. patents by keyword, inventor, assignee, CPC code, date, and type |
ip_eu_patents | EPO Open Patent Services | eu-patent-search | Search European patents by title, applicant, publication number, country |
ip_eu_trademarks | EUIPO | eu-trademark-search | Search EU trademarks by verbal element, Nice class, status, applicant |
ip_global_patent_search | USPTO + EPO | global-patent-search | Cross-jurisdictional search with CPC overlap and common filer analysis |
ip_list_sources | — | Free | List all available tools with credential requirements |
Tool parameters
ip_us_patents
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
keyword | string | No | — | Full-text search in patent titles |
inventor | string | No | — | Inventor last name |
assignee | string | No | — | Assignee organization name (e.g. "Apple", "Samsung") |
date_from | string | No | — | Grant date start, YYYY-MM-DD |
date_to | string | No | — | Grant date end, YYYY-MM-DD |
cpc_code | string | No | — | CPC classification prefix (e.g. "G06F" for computing, "H01M" for batteries) |
patent_type | enum | No | — | utility, design, plant, or reissue |
sort_by | enum | No | patent_date | patent_date, patent_id, or patent_num_us_patents_cited |
api_key | string | No | — | Free PatentsView API key from patentsview.org/apis |
max_results | number | No | 50 | Max results, 1–500 |
ip_eu_patents
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
title_keywords | string | No | — | Keywords in patent title |
applicant | string | No | — | Applicant/assignee name |
inventor | string | No | — | Inventor name |
publication_number | string | No | — | Direct lookup by number (e.g. "EP1234567") |
publication_country | string | No | — | Country code: EP, DE, FR, GB, etc. |
date_from | string | No | — | Publication date start (YYYY or YYYYMMDD) |
date_to | string | No | — | Publication date end |
query | string | No | — | Raw CQL query — overrides all other search fields |
consumer_key | string | No | — | EPO OPS consumer key from developers.epo.org |
consumer_secret | string | No | — | EPO OPS consumer secret |
max_results | number | No | 50 | Max results, 1–500 |
ip_eu_trademarks
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
query | string | Yes | — | Trademark verbal element / name to search |
nice_class | string | No | — | Nice class numbers, comma-separated 1–45 (e.g. "9,35,42") |
status | enum | No | — | REGISTERED, EXPIRED, CANCELLED, REFUSED, FILED, OPPOSED, or WITHDRAWN |
applicant_name | string | No | — | Applicant name, partial match supported |
mark_feature | enum | No | — | WORD, FIGURATIVE, SHAPE_3D, COLOUR, SOUND, POSITION, PATTERN, MOTION, MULTIMEDIA, or OTHER |
date_from | string | No | — | Application date start, YYYY-MM-DD |
date_to | string | No | — | Application date end |
client_id | string | No | — | EUIPO developer portal client ID |
client_secret | string | No | — | EUIPO developer portal client secret |
max_results | number | No | 25 | Max results, 1–200 |
ip_global_patent_search
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
keyword | string | Yes | — | Technology keyword or phrase |
inventor | string | No | — | Inventor name to search in both databases |
assignee | string | No | — | Company name to search in both databases |
date_from | string | No | — | Start date (YYYY-MM-DD for USPTO, YYYY for EPO) |
date_to | string | No | — | End date |
uspto_api_key | string | No | — | PatentsView API key |
epo_consumer_key | string | No | — | EPO OPS consumer key |
epo_consumer_secret | string | No | — | EPO OPS consumer secret |
max_per_source | number | No | 25 | Max results per database, 1–100 |
API credentials setup
All three data sources provide free API credentials. No paid subscriptions are required.
| Source | Registration URL | What you get |
|---|---|---|
| USPTO PatentsView | patentsview.org/apis | Free API key, no usage limits stated |
| EPO Open Patent Services | developers.epo.org | Consumer key + secret, 4 MB/week free tier |
| EUIPO developer portal | dev.euipo.europa.eu | Client ID + secret for trademark API |
Without credentials, each tool detects the missing key and returns setup instructions rather than failing with an error.
Prompt examples
Once connected, ask your AI assistant:
Search US patents assigned to "QuantumDyne Systems" in the battery technology CPC code H01M filed after 2022.
Use my PatentsView key: pv_abc123
Search EU trademarks for the name "NEXAFLOW" in Nice classes 9 and 42, registered status only.
Run a global patent landscape search for "solid state electrolyte" across both USPTO and EPO,
max 30 results per source. My API keys: [USPTO: pv_abc123, EPO key: abc, EPO secret: xyz]
Output examples
ip_us_patents response
{
"total": 3,
"uniqueAssignees": 2,
"totalCitations": 94,
"data": [
{
"patentId": "US11482723",
"title": "Solid-state electrolyte with ceramic nanoparticle composite",
"abstract": "A solid-state electrolyte composition comprising ceramic nanoparticles dispersed in a polymer matrix...",
"grantDate": "2022-10-25",
"patentType": "utility",
"inventors": "Chen, Wei; Nakamura, Taro; Liu, Fang",
"assignees": "QuantumDyne Systems Inc.",
"citationCount": 47,
"cpcCodes": ["H01M10/0525", "H01M10/0562", "C01B33/12"],
"patentNumber": "11482723"
},
{
"patentId": "US11398640",
"title": "Method for manufacturing thin-film solid electrolyte layers",
"abstract": "A scalable deposition method for producing uniform thin-film solid electrolyte layers...",
"grantDate": "2022-07-26",
"patentType": "utility",
"inventors": "Rodriguez, Maria; Patel, Arun",
"assignees": "Pinnacle Energy Technologies LLC",
"citationCount": 31,
"cpcCodes": ["H01M10/0525", "H01M4/38"],
"patentNumber": "11398640"
}
]
}
ip_eu_patents response
{
"total": 2,
"countries": ["EP", "DE"],
"data": [
{
"publicationNumber": "EP4123456",
"title": "Festkörperelektrolyt mit keramischen Nanopartikeln",
"applicants": ["QuantumDyne Systems Inc."],
"inventors": ["Chen, Wei", "Nakamura, Taro"],
"publicationDate": "20230315",
"country": "EP",
"cpcClassifications": ["H01M10/0525", "H01M10/0562"],
"ipcClassifications": ["H01M10/052"],
"familyId": "74812349"
}
]
}
ip_eu_trademarks response
{
"total": 2,
"byStatus": {
"REGISTERED": 1,
"FILED": 1
},
"data": [
{
"applicationNumber": "018934561",
"verbalElement": "NEXAFLOW",
"status": "REGISTERED",
"applicantName": "Nexaflow Technologies GmbH",
"niceClasses": ["9", "35", "42"],
"goodsAndServices": "Software for workflow automation (9); Business process consulting (35); Software as a service (42)",
"applicationDate": "2021-06-14",
"registrationDate": "2022-03-08",
"markFeature": "WORD",
"imageUrl": null
}
]
}
ip_global_patent_search response
{
"keyword": "solid state electrolyte",
"sources": ["USPTO (PatentsView)", "EPO (Open Patent Services)"],
"usPatents": {
"total": 25,
"hasData": true,
"results": ["..."]
},
"euPatents": {
"total": 18,
"hasData": true,
"results": ["..."]
},
"crossJurisdiction": {
"sharedCpcCodes": ["H01M", "C01B"],
"commonFilers": 3,
"note": "2 CPC code group(s) appear in both US and EU results, suggesting overlapping technology areas"
}
}
How much does it cost to search patents and trademarks?
The IP Patent & Trademark MCP Server uses pay-per-event pricing at $0.045 per tool call. The ip_global_patent_search tool counts as one event despite querying two databases in parallel. Standby mode keeps the server warm between calls at no charge.
| Scenario | Tool calls | Cost per call | Total cost |
|---|---|---|---|
| Quick test | 1 | $0.045 | $0.045 |
| Small research session | 10 | $0.045 | $0.45 |
| Daily IP monitoring workflow | 50 | $0.045 | $2.25 |
| Weekly landscape analysis | 200 | $0.045 | $9.00 |
| Enterprise continuous monitoring | 1,000 | $0.045 | $45.00 |
You can set a maximum spending limit per run to control costs. The actor stops when your budget is reached.
Apify's free tier includes $5 of monthly platform credits — enough for approximately 111 tool calls at no cost. Compare this to commercial patent intelligence platforms like PatSnap or Derwent Innovation at $500–2,000/month; for most AI-assisted IP workflows, monthly spend with this server falls under $10.
How IP Patent & Trademark MCP Server works
Architecture
The server runs as an Express.js HTTP server on Apify's standby infrastructure, implementing the Streamable HTTP MCP transport from @modelcontextprotocol/sdk. Each POST request to /mcp instantiates a fresh McpServer, connects it to a new StreamableHTTPServerTransport, processes the request, then tears the instance down — ensuring full request isolation with no shared state between calls.
Tool registration uses Zod schemas for input validation. All 5 tools are registered on every request cycle, with parameter validation happening before any actor is invoked.
Data retrieval via actor orchestration
Each MCP tool maps to one or more Apify actors in the ryanclinton namespace. The actor-client.ts module uses apify-client to call each actor synchronously with a 120-second waitSecs timeout and a fixed 256 MB memory allocation. Results are fetched from the run's default dataset via listItems().
The three underlying actors called are:
ryanclinton/patent-search— wraps the PatentsView REST API (/api/v1/patents/query)ryanclinton/epo-patent-search— wraps the EPO OPS REST API with OAuth2 token exchangeryanclinton/euipo-trademark-search— wraps the EUIPO EUIPOTM API
Cross-jurisdictional analysis logic
The ip_global_patent_search tool runs the USPTO and EPO actors with Promise.all() for parallel execution. After both complete, the tool:
- Extracts CPC codes from U.S. results, taking the first 4 characters (the subclass level) as group identifiers
- Does the same for EU results using
cpcClassifications - Computes the intersection to find shared CPC subclass groups
- Compares assignee names (lowercased) between U.S. and EU results using substring matching to identify companies filing in both jurisdictions
Results are returned as crossJurisdiction.sharedCpcCodes (the overlapping subclass groups) and crossJurisdiction.commonFilers (the count of cross-border assignees).
Billing and spend control
Each tool registers its own charge event name (e.g. us-patent-search, eu-trademark-search). Before executing, the tool calls Actor.charge({ eventName }). If chargeResult.eventChargeLimitReached is true, the tool returns immediately with an error rather than consuming quota, preventing overspend.
Tips for best results
-
Register all three API credentials before your first session. PatentsView, EPO OPS, and EUIPO all offer free keys with same-day approval. Store them as inputs you pass to the tool, or ask your AI client to request them at session start.
-
Use CPC prefix codes for technology landscape searches. The 4-character subclass (e.g.
H01Mfor electrochemical energy storage) returns broad coverage; a longer code likeH01M10/0525narrows to a specific technology. Start broad, then narrow based on results. -
Use
ip_global_patent_searchfor initial scoping. It costs the same as a single-database search but returns twice the data and highlights where CPC codes overlap. Switch toip_us_patentsorip_eu_patentsindividually only when you need source-specific filters likepatent_type(USPTO) or raw CQL (EPO). -
Filter EU trademarks by REGISTERED status for clearance checks. Including EXPIRED and CANCELLED marks adds noise to clearance searches. Set
status: "REGISTERED"unless you specifically need the history. -
Sort US patents by
patent_num_us_patents_citedto find foundational IP. High citation counts indicate seminal patents with broad industry influence. Use this sort when assessing a technology's IP landscape quality, not just volume. -
Combine with EUIPO Trademark Search directly for batched trademark monitoring work that doesn't require an interactive AI session. The standalone actor supports scheduling and CSV output.
-
For U.S. trademark coverage, complement with a USPTO TESS search. This server covers EU trademarks via EUIPO only. U.S. federal trademark registrations are held in the USPTO TESS database, which is a separate system.
Combine with other Apify actors
| Actor | How to combine |
|---|---|
| EPO Patent Search | Run standalone for large batch European patent downloads outside an AI session, with scheduled monitoring of new EP filings |
| EUIPO Trademark Search | Schedule recurring trademark clearance scans for a brand portfolio and export to Google Sheets |
| Website Tech Stack Detector | Identify technology vendors used by a company, then cross-reference their patent portfolio using ip_us_patents to assess IP depth |
| Company Deep Research | Combine corporate intelligence reports with patent landscape data for M&A due diligence workflows |
| WHOIS Domain Lookup | Verify domain registration ownership against trademark applicant names to identify brand registration conflicts |
| HubSpot Lead Pusher | Push companies identified through patent assignee searches into a CRM as technology-qualified prospects |
| Website Content to Markdown | Convert patent attorney or IP firm websites to markdown for LLM ingestion alongside patent data for combined analysis |
Limitations
- All three data sources require free API credentials. Without them, each tool returns setup instructions rather than data. Registration is free and typically same-day, but cannot be skipped.
- PatentsView covers U.S. patents only. Design patents and plant patents are included, but PCT applications not yet granted in the U.S. are not available via PatentsView.
- EPO OPS free tier has a bandwidth cap. The free EPO OPS tier limits consumption to approximately 4 MB per week. Large queries returning 500 results may exhaust this quickly. Monitor your EPO usage at developers.epo.org.
- EUIPO covers EU-wide (EUTM) trademarks, not national registrations. Trademarks filed in Germany, France, or the UK separately are not visible through EUIPO. For national coverage, each country's trademark office must be queried individually.
- Cross-jurisdictional deduplication is approximate. The
ip_global_patent_searchtool identifies overlaps using 4-character CPC subclass codes and lowercased assignee substring matching, which may miss patent families with inconsistent naming or fail to detect all duplicates across jurisdictions. - Patent data has a publication lag. PatentsView and EPO OPS data may lag behind real-time filings by days to weeks. Freshly filed applications may not yet appear.
- The server does not persist sessions. Each MCP request is stateless. Pagination across large result sets requires repeated calls with explicit offset parameters handled by the underlying actors.
- This server is not a substitute for professional IP counsel. Patent and trademark searches here are informational. Freedom-to-operate and clearance decisions should be reviewed by a qualified IP attorney.
Integrations
- Zapier — trigger patent landscape searches from Zaps that monitor competitor filings or news mentions
- Make — build automated IP monitoring workflows that push new trademark conflicts into a project management tool
- Google Sheets — export patent or trademark search results to a shared team spreadsheet for collaborative review
- Apify API — call the MCP server programmatically from Python or JavaScript pipelines without an AI client
- Webhooks — receive notifications when a scheduled IP monitoring run completes or encounters an error
- LangChain / LlamaIndex — use patent data retrieved via this server as a retrieval source for RAG pipelines focused on technology or legal research
Troubleshooting
Tool returns setup instructions instead of patent data. Each data source requires a free API key passed as a tool parameter. Check that you are passing api_key for ip_us_patents, consumer_key and consumer_secret for ip_eu_patents, and client_id and client_secret for ip_eu_trademarks. Call ip_list_sources first to confirm credential requirements for each tool.
ip_eu_patents returns empty results despite a valid query. EPO OPS uses CQL (Contextual Query Language) syntax. If using the query parameter directly, ensure CQL operators are correct (e.g. ta="solid state battery" for title-or-abstract search). Alternatively, use the structured title_keywords, applicant, and inventor fields and let the actor construct the query.
ip_global_patent_search shows data for one source but not the other. The tool runs both actors in parallel and filters out items containing an info or message field (the credential-missing sentinel). If one source shows noApiKey: true in the response, add the missing credential for that source and retry.
EPO OPS calls fail intermittently. The EPO OPS free tier has a weekly bandwidth limit of approximately 4 MB. If you have run many large queries in the same week, you may have hit the cap. Check your quota at developers.epo.org. Reducing max_results per query extends your weekly budget.
Long response times on the first call after an idle period. Apify standby mode keeps the server warm, but if the server has been idle for an extended period it may restart. The first request after restart incurs a cold-start delay of a few seconds. Subsequent requests within the session are fast.
Responsible use
- This server only accesses publicly available IP data via official API endpoints operated by USPTO, EPO, and EUIPO.
- Patent and trademark databases are public records. Use of the data is subject to the terms of service of each respective source.
- Do not use patent or trademark data to harass individuals, engage in bad-faith IP filings, or circumvent legitimate IP rights.
- For guidance on web scraping and data access legality, see Apify's guide.
FAQ
How do I search patents by CPC classification code?
Use ip_us_patents with the cpc_code parameter. Pass a 4-character subclass prefix like H01M (batteries) or G06F (computing) for broad technology coverage, or a full CPC code like H01M10/0525 to narrow to a specific technology. For EPO, pass a CQL query like cl="H01M10/0525" using the query parameter in ip_eu_patents.
How many patent results can I retrieve in one call?
ip_us_patents and ip_eu_patents support up to 500 results per query. ip_eu_trademarks supports up to 200 results. ip_global_patent_search supports up to 100 results per source (200 total). For comprehensive landscape analysis requiring more results, run multiple queries with date ranges to segment the result set.
Does ip_global_patent_search find the same patent filed in both the US and Europe? It attempts to detect overlap using shared CPC subclass codes (first 4 characters) and common assignee name matching. This identifies the same technology cluster and filer in both jurisdictions, but it is not full patent family deduplication. For true family-level deduplication, use the EPO's INPADOC family data available via the raw CQL query interface.
Is it legal to access USPTO, EPO, and EUIPO data via API? Yes. All three sources are official government or intergovernmental patent and trademark offices that provide public APIs specifically for this purpose. PatentsView, EPO Open Patent Services, and the EUIPO API are official programmatic access channels, not scraped data.
Do I need a paid subscription to use any of the data sources? No. PatentsView, EPO OPS, and EUIPO all offer free API credentials. EPO OPS has a free bandwidth limit of approximately 4 MB per week on the free tier. PatentsView and EUIPO do not publish explicit rate limits for the free tier. Heavy commercial use may warrant registering for higher-tier access with EPO.
How is this different from PatSnap or Derwent Innovation? PatSnap and Derwent Innovation are full patent intelligence platforms with analytics dashboards, citation mapping, and portfolio tracking — suited to dedicated IP teams with large analysis budgets ($500–2,000+/month). This server provides the same raw patent data access at pay-per-use pricing ($0.045/query), accessible directly from your AI assistant without switching to a separate application. It is best suited to teams that need IP data as one input among many in an AI-assisted research workflow.
Can I search US federal trademark registrations?
The ip_eu_trademarks tool covers EU-wide trademark registrations (EUTM) via EUIPO only. U.S. federal trademark registrations are held in the USPTO TESS database, which is a separate system not currently covered by this server. For US trademark clearance, use the USPTO TESS search directly at tmsearch.uspto.gov.
How often is the patent data updated? PatentsView updates its database weekly from USPTO grant data. EPO OPS data is updated continuously but publication of new applications typically happens approximately 18 months after filing. EUIPO trademark data reflects the official EUIPO register, which updates as applications are processed.
Can I schedule recurring patent monitoring with this server?
The MCP server is designed for interactive AI sessions. For scheduled recurring searches (e.g. weekly new filings in a technology class), use the standalone ryanclinton/patent-search, ryanclinton/epo-patent-search, or ryanclinton/euipo-trademark-search actors directly via Apify scheduling, which supports cron-based scheduling and email notifications.
What happens if one data source is unavailable when I call ip_global_patent_search?
The runActor function in actor-client.ts wraps each actor call in a try/catch. If one source fails, it returns an empty array for that source rather than failing the entire request. The response will show hasData: false for the failed source, allowing you to see results from the other source while identifying which one needs attention.
Can I use this MCP server with LangChain or other AI frameworks?
Yes. Any framework that supports the Streamable HTTP MCP transport can connect to this server. For LangChain and LlamaIndex, use the Apify integration documented at docs.apify.com/platform/integrations. The server endpoint is https://ip-patent-trademark-mcp.apify.actor/mcp.
Help us improve
If you encounter issues, you can help us debug faster by enabling run sharing in your Apify account:
- Go to Account Settings > Privacy
- Enable Share runs with public Actor creators
This lets us see your run details when something goes wrong, so we can fix issues faster. Your data is only visible to the actor developer, not publicly.
Support
Found a bug or have a feature request? Open an issue in the Issues tab on this actor's page. For custom solutions or enterprise integrations, reach out through the Apify platform.
How it works
Configure
Set your parameters in the Apify Console or pass them via API.
Run
Click Start, trigger via API, webhook, or set up a schedule.
Get results
Download as JSON, CSV, or Excel. Integrate with 1,000+ apps.
Use cases
Sales Teams
Build targeted lead lists with verified contact data.
Marketing
Research competitors and identify outreach opportunities.
Data Teams
Automate data collection pipelines with scheduled runs.
Developers
Integrate via REST API or use as an MCP tool in AI workflows.
Related actors
Bulk Email Verifier
Verify email deliverability at scale. MX record validation, SMTP mailbox checks, disposable and role-based detection, catch-all flagging, and confidence scoring. No external API costs.
GitHub Repository Search
Search GitHub repositories by keyword, language, topic, stars, forks. Sort by stars, forks, or recently updated. Returns metadata, topics, license, owner info, URLs. Free API, optional token for higher limits.
Website Content to Markdown
Convert any website to clean Markdown for RAG pipelines, LLM training, and AI apps. Crawls pages, strips boilerplate, preserves headings, tables, and code blocks. GFM support.
Website Tech Stack Detector
Detect 100+ web technologies on any website. Identifies CMS, frameworks, analytics, marketing tools, chat widgets, CDNs, payment systems, hosting, and more. Batch-analyze multiple sites with version detection and confidence scoring.
Ready to try IP Patent & Trademark MCP Server?
Start for free on Apify. No credit card required.
Open on Apify Store