home / mcp / patent mcp server

Patent MCP Server

Provides patent data queries and analytics, enabling fuzzy search, patent search, and statistics for IP insights.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "handaas-patent-mcp-server": {
      "url": "http://127.0.0.1:8000/mcp",
      "headers": {
        "SECRET_ID": "your_secret_id",
        "SECRET_KEY": "your_secret_key",
        "INTEGRATOR_ID": "your_integrator_id"
      }
    }
  }
}

You can access Patent MCP Server to query comprehensive patent data and perform analytics for IP intelligence. Run locally or connect through an MCP client to search patents, analyze patent statistics, and explore fuzzy enterprise keyword queries.

How to use

You use an MCP client to connect to Patent MCP Server either via an HTTP endpoint or by running the server locally in stdio mode. Once connected, you can perform three main actions: perform fuzzy searches for enterprise keywords, search patent information, and generate statistics about a company’s patent activity. Use the HTTP connection for quick remote access or run the local stdio server to host the MCP locally and connect through the 127.0.0.1:8000/mcp endpoint.

How to install

Prerequisites: Python 3.10+ and a working shell. You will also install Python dependencies and configure environment variables before starting the server.

git clone https://github.com/handaas/patent-mcp-server
cd patent-mcp-server
```

```bash
python -m venv mcp_env && source mcp_env/bin/activate
pip install -r requirements.txt
```

```bash
cp .env.example .env
```

Edit the .env file to set:

```
INTEGRATOR_ID=your_integrator_id
SECRET_ID=your_secret_id
SECRET_KEY=your_secret_key
```

Then start the local MCP server in streamable-http mode:

```
python server/mcp_server.py streamable-http
```

The service will be available at http://localhost:8000/mcp.

Configuration and connections

You can connect with either an HTTP MCP endpoint or a local stdio MPC command. Use the provided examples to configure your MCP client or integration tool.

{
  "mcpServers": {
    "patent_mcp": {
      "type": "http",
      "url": "http://127.0.0.1:8000/mcp",
      "args": []
    }
  }
}

{
  "mcpServers": {
    "patent_mcp_stdio": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "mcp", "run", "{workdir}/server/mcp_server.py"],
      "env": {
        "PATH": "{workdir}/mcp_env/bin:$PATH",
        "PYTHONPATH": "{workdir}/mcp_env",
        "INTEGRATOR_ID": "your_integrator_id",
        "SECRET_ID": "your_secret_id",
        "SECRET_KEY": "your_secret_key"
      }
    }
  }
}

Notes and best practices

- One page of results is limited to 50 entries for each query. Use pagination with pageIndex and pageSize to navigate larger result sets. - If you need to know the full enterprise name for certain interfaces, first run a fuzzy enterprise search to retrieve the complete name. - The available patent types include invention, utility model, and design, and you can filter by type when searching patents.

Common use cases

- Patent代理机构: quickly assess a competitor’s patent布局 by querying the latest patents and statistics. - 研发人员: identify existing patents in a field to avoid potential infringement. - 企业技术创新: perform intelligence analysis to gauge market and technology trends. - 投资分析: evaluate a company’s patent portfolio and innovation capacity. - 法律诉讼: assess infringement risk based on patent activity. - 科技项目评估: understand patent distribution across technology areas.

Available tools

patent_bigdata_fuzzy_search

Fuzzy search for enterprise keywords to find matching companies based on partial names, person names, brands, products, or roles.

patent_bigdata_patent_search

Search patent information by patent name, application number, applicant, or agent, with optional type filtering.

patent_bigdata_patent_stats

Analyze a company’s patent portfolio, including status distribution and application/grant trends.