PatSnap Patent & Literature Search MCP

Skill by ara.so — MCP Skills collection.

Overview

PatSnap Patent & Literature Search MCP provides AI agents with direct access to 200M+ patents across 170+ jurisdictions (USPTO, EPO, WIPO) and 216M+ scientific papers. Search using natural language queries without learning Boolean syntax. Supports fusion search (patents + literature), semantic search, keyword/BM25 text search, and precise filtering by assignee, inventor, IPC class, legal status, jurisdiction, date range, and citations.

Installation

Claude Desktop

claude mcp add --transport http search-tool \
  "https://connect.patsnap.com/2b0355/logic-mcp?apikey=${PATSNAP_API_KEY}"

Cursor or Windsurf

Add to MCP configuration file (~/Library/Application Support/Cursor/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json or equivalent):

{
  "mcpServers": {
    "patsnap_patent_literature": {
      "url": "https://connect.patsnap.com/2b0355/logic-mcp?apikey=${PATSNAP_API_KEY}",
      "type": "streamableHttp"
    }
  }
}

Docker (stdio bridge)

docker run --rm -i \
  -e PATSNAP_API_KEY=${PATSNAP_API_KEY} \
  ghcr.io/patsnap/patent-literature-search-mcp:latest

MCP client configuration for Docker:

{
  "mcpServers": {
    "patsnap_patent_literature": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e",
        "PATSNAP_API_KEY",
        "ghcr.io/patsnap/patent-literature-search-mcp:latest"
      ],
      "env": {
        "PATSNAP_API_KEY": "${PATSNAP_API_KEY}"
      }
    }
  }
}

API Key

1. Register at Patsnap Open Platform
2. Navigate to Patent & Literature Search MCP page
3. Generate API key (10,000 free credits for new accounts)
4. Set environment variable: export PATSNAP_API_KEY=your_key_here

Available Tools

patsnap_search

Search patents or scientific literature with natural language or structured queries.

Parameters:

• source (required): "patent" or "literature"
• query (required): Natural language search query or keywords
• filters (optional): Object with filter criteria
  • assignee: Company/organization name
  • inventor: Inventor name
  • ipc_class: International Patent Classification code
  • legal_status: Patent legal status
  • jurisdiction: Country/region code (e.g., "US", "EP", "CN")
  • date_from, date_to: Date range (ISO 8601 format)
  • min_citations: Minimum citation count
• page (optional): Page number (default: 1)
• page_size (optional): Results per page (default: 10, max: 50)

Example: Patent search with filters

{
  "source": "patent",
  "query": "solid state battery lithium metal anode",
  "filters": {
    "assignee": "Toyota",
    "date_from": "2020-01-01",
    "jurisdiction": "US",
    "legal_status": "granted"
  },
  "page": 1,
  "page_size": 20
}

Example: Literature search

{
  "source": "literature",
  "query": "GLP-1 receptor agonist cardiovascular outcomes",
  "filters": {
    "date_from": "2022-01-01",
    "min_citations": 10
  },
  "page_size": 15
}

Example: IPC classification search

{
  "source": "patent",
  "query": "neural network image recognition",
  "filters": {
    "ipc_class": "G06N3/08",
    "date_from": "2023-01-01"
  }
}

patsnap_fetch

Retrieve detailed patent or literature records as Markdown.

Parameters:

• urls (optional): Array of result URLs from patsnap_search
• publication_numbers (optional): Array of patent publication numbers (e.g., ["US10123456B2", "EP3456789A1"])

Example: Fetch by URLs

{
  "urls": [
    "https://analytics.zhihuiya.com/patent-view?patentId=abc123",
    "https://analytics.zhihuiya.com/patent-view?patentId=def456"
  ]
}

Example: Fetch by publication numbers

{
  "publication_numbers": ["US11234567B2", "WO2023012345A1"]
}

Common Usage Patterns

Prior Art Search

Search for existing patents before filing:

// Step 1: Search patents
{
  "source": "patent",
  "query": "augmented reality waveguide display optical combiner",
  "filters": {
    "date_to": "2023-06-01",
    "jurisdiction": "US"
  },
  "page_size": 50
}

// Step 2: Fetch detailed records of top results
{
  "urls": ["<url1>", "<url2>", "<url3>"]
}

Technology Landscape Analysis

Map patent activity by assignee and technology area:

// Search broad technology area
{
  "source": "patent",
  "query": "electric vehicle battery thermal management",
  "filters": {
    "date_from": "2020-01-01"
  },
  "page_size": 50
}

// Follow up with assignee-specific searches
{
  "source": "patent",
  "query": "battery thermal management",
  "filters": {
    "assignee": "Tesla",
    "date_from": "2020-01-01"
  }
}

Fusion Search (Patents + Literature)

Combine patent and academic research:

// Search patents
{
  "source": "patent",
  "query": "CRISPR base editing adenine deaminase",
  "filters": {
    "date_from": "2021-01-01",
    "legal_status": "granted"
  }
}

// Search literature on same topic
{
  "source": "literature",
  "query": "CRISPR base editing adenine deaminase efficiency",
  "filters": {
    "date_from": "2021-01-01",
    "min_citations": 5
  }
}

Patent Expiration Monitoring

Find patents expiring soon:

{
  "source": "patent",
  "query": "GLP-1 receptor agonist diabetes treatment",
  "filters": {
    "legal_status": "granted",
    "date_to": "2008-12-31" // Filed ~20 years ago
  },
  "page_size": 30
}

Inventor/Assignee Tracking

Monitor specific organizations or inventors:

{
  "source": "patent",
  "query": "quantum computing error correction",
  "filters": {
    "assignee": "IBM",
    "date_from": "2023-01-01"
  },
  "page_size": 25
}

Response Structure

Search Response

{
  "total": 1234,
  "page": 1,
  "page_size": 10,
  "results": [
    {
      "title": "Solid-state battery with lithium metal anode",
      "url": "https://analytics.zhihuiya.com/patent-view?patentId=...",
      "publication_number": "US11234567B2",
      "assignee": "Toyota Motor Corporation",
      "filing_date": "2021-03-15",
      "publication_date": "2023-01-24",
      "abstract": "A solid-state battery comprising...",
      "ipc_classes": ["H01M10/0562", "H01M10/0525"],
      "citations": 12
    }
  ]
}

Fetch Response

Returns Markdown-formatted detailed records including:

• Patents: Bibliographic data, claims, description, drawings, citations, legal status, patent family
• Literature: Title, authors, journal, DOI, abstract, citations, publication info

Configuration

Environment Variables

• PATSNAP_API_KEY (required): API key from Patsnap Open Platform

Docker Build (for self-hosting)

FROM node:20-alpine
WORKDIR /app
COPY package*.json ./
RUN npm ci --production
COPY src ./src
CMD ["node", "src/index.js"]

Build and run:

docker build -t patsnap-search-mcp .
docker run --rm -i -e PATSNAP_API_KEY=${PATSNAP_API_KEY} patsnap-search-mcp

Troubleshooting

Connection Issues

Problem: MCP server not appearing in Claude/Cursor

Solution:

• Verify API key is set: echo $PATSNAP_API_KEY
• Check MCP configuration file syntax (valid JSON)
• Restart IDE/Claude Desktop after configuration changes
• Test with: claude mcp list (Claude Desktop)

Search Returns No Results

Problem: Query returns zero results

Solution:

• Simplify query (remove overly specific terms)
• Remove or relax filters (date range, jurisdiction)
• Try different keyword combinations
• Use broader IPC classes (e.g., "H01M10" instead of "H01M10/0562")
• Check spelling of assignee/inventor names

API Rate Limits

Problem: Request fails with rate limit error

Solution:

• Free tier: 10,000 credits (check usage on Patsnap Open Platform)
• Reduce page_size to consume fewer credits
• Implement exponential backoff for repeated searches
• Upgrade account for higher limits

Invalid Publication Number

Problem: patsnap_fetch fails with publication number

Solution:

• Use standardized format: US11234567B2 (country code + number + kind code)
• Common formats: US10123456B2, EP3456789A1, WO2023012345A1
• Extract from publication_number field in search results

Docker Container Exits Immediately

Problem: Container starts but exits without output

Solution:

• Ensure -i flag is present: docker run --rm -i
• Verify MCP client sends proper stdio handshake
• Check container logs: docker logs <container_id>
• Confirm PATSNAP_API_KEY is passed: docker run -e PATSNAP_API_KEY

Best Practices

1. Start broad, then filter: Begin with general queries, analyze results, then add filters
2. Combine sources: Use both patent and literature searches for comprehensive research
3. Iterate on queries: Refine based on initial results (synonyms, related terms)
4. Use fetch selectively: Only fetch full records for most relevant results (saves credits)
5. Leverage IPC codes: Use classification codes for precise technical searching
6. Date ranges: Narrow by date to focus on recent innovations or identify expiring patents
7. Citation counts: Filter by citations to find influential work (especially for literature)

Example Workflows

Competitive Intelligence

// 1. Find all patents by competitor
const competitorPatents = await patsnap_search({
  source: "patent",
  query: "autonomous vehicle perception lidar",
  filters: { assignee: "Waymo" },
  page_size: 50
});

// 2. Analyze their recent focus areas
const recentPatents = await patsnap_search({
  source: "patent",
  query: "autonomous vehicle",
  filters: {
    assignee: "Waymo",
    date_from: "2023-01-01"
  }
});

// 3. Fetch details for key patents
const details = await patsnap_fetch({
  urls: competitorPatents.results.slice(0, 5).map(r => r.url)
});

Freedom to Operate (FTO)

// 1. Search active patents in target jurisdiction
const activePatents = await patsnap_search({
  source: "patent",
  query: "mRNA vaccine lipid nanoparticle formulation",
  filters: {
    jurisdiction: "US",
    legal_status: "granted",
    date_from: "2003-01-01" // Within 20-year term
  },
  page_size: 50
});

// 2. Fetch full claims for analysis
const claims = await patsnap_fetch({
  publication_numbers: activePatents.results.map(r => r.publication_number)
});

Resources

• Patsnap Open Platform
• API Documentation
• Eureka AI Assistant
• GitHub Repository
• All Patsnap MCP Servers