Information Big Data Service

This MCP service provides enterprise keyword search, global information search, enterprise news, industry news, topic tracking, and enterprise dynamics monitoring capabilities, helping users conduct industry research, hot topic tracking, enterprise public opinion analysis, and brand reputation management.

Main Features

• 🏢 Enterprise abbreviation and keyword search

• 🔍 Global information keyword search

• 🏢 Enterprise news and public opinion query

• 🏭 Industry and track information search

• 🎯 Policy, technology, brand, and event topic tracking

• 📊 Enterprise information details and sentiment statistics monitoring

Related MCP server: industry-chain-mcp-server

Service Design Notes

• The service provides 6 Tools based on actual business scenarios, rather than laying out tools by the number of upstream APIs.

• When users only provide an enterprise abbreviation, first use information_enterprise_search to obtain the full enterprise name or stable ID.

• The search, industry, and topic Tools reuse the same global information search Product ID, but retain different scenario semantics.

• Upstream information search results are paginated by the MCP layer via pageIndex and pageSize, with a default of 20 items per page and a maximum of 50.

• Search results do not return long content by default; after setting includeContent=true, the length of each content item can be limited via contentMaxChars.

• information_monitor selects details or statistics via view, accessing only one Product ID at a time.

• The business outer layer of all paginated responses only contains total and resultList.

Environment Requirements

• Python 3.10+

• Dependencies: python-dotenv, requests, mcp

Local Quick Start

1. Enter the Project Directory

cd information-mcp-server

2. Create a Virtual Environment and Install Dependencies

python3 -m venv mcp_env
source mcp_env/bin/activate
pip install -r requirements.txt

3. Configure Environment Variables

Copy the environment variable template:

cp .env.example .env

Edit the .env file:

INTEGRATOR_ID=your_integrator_id
SECRET_ID=your_secret_id
SECRET_KEY=your_secret_key
HANDAAS_REQUEST_TIMEOUT=30

HANDAAS_REQUEST_TIMEOUT is an optional configuration, in seconds, with a default value of 30.

4. Start the Streamable HTTP Service

python server/mcp_server.py streamable-http

The default service address is http://localhost:8000/mcp.

You can also use the startup script:

./start_mcp_server.sh streamable-http

Supports three startup modes: stdio, sse, and streamable-http.

5. Cursor / Cherry Studio MCP Configuration

{
  "mcpServers": {
    "information-mcp-server": {
      "type": "streamableHttp",
      "url": "http://127.0.0.1:8000/mcp"
    }
  }
}

STDIO Version Installation and Deployment

Replace {workdir} with the absolute path of information-mcp-server:

{
  "mcpServers": {
    "information-mcp-server": {
      "command": "{workdir}/mcp_env/bin/python",
      "args": [
        "{workdir}/server/mcp_server.py",
        "stdio"
      ]
    }
  }
}

INTEGRATOR_ID, SECRET_ID, and SECRET_KEY need to be obtained after logging into HandaaS to register and activate the connector. Real credentials should only be stored in the local .env or deployment secrets.

Available Tools and Product IDs

1. information_enterprise_search

Function: Search for candidate enterprises by enterprise abbreviation, brand, product, or other keywords.

Main parameters: matchKeyword is required; pageIndex defaults to 1; pageSize defaults to 10, maximum 50.

Returns: Candidate enterprises total/resultList. After confirming the enterprise, use the full enterprise name, enterprise ID, or unified social credit code for the enterprise news and monitoring Tools.

2. information_search

Function: Search news information titles and content by keyword.

Main parameters:

• matchKeyword (required): Search keyword.

• pubDateBegin, pubDateEnd (optional): Publication time start and end conditions.

• pageIndex (optional): MCP return page number, starting from 1.

• pageSize (optional): Number of items per page, default 20, maximum 50.

• includeContent (optional): Whether to return the information content, default false.

• contentMaxChars (optional): Maximum number of characters to keep per item when returning content, default 1000, range 100–5000.

Returns: total and resultList; results may include title, source, publication time, information link, and content truncated as needed.

3. information_company_news

Function: Query the news and public opinion details of a specified enterprise.

Main parameters:

• matchKeyword (required): Enterprise name, enterprise ID, registration number, or unified social credit code.

• keywordType (optional): Enterprise identifier type, supports name, nameId, regNumber, socialCreditCode.

• pageIndex (optional): Page number, starting from 1.

• pageSize (optional): Number of items per page, default 50, maximum 50.

• sentimentLabel (optional): 0 negative, 1 positive, 2 neutral, 3 unknown.

Returns: total and resultList; results may include news summary, link, source, title, publication time, related enterprises, and sentiment label.

4. information_industry_news

Function: Search information by industry, track, or sector keywords for industry research and market dynamics analysis.

Main parameters:

• matchKeyword (required): Industry, track, or sector keyword.

• pubDateBegin, pubDateEnd (optional): Publication time range.

• pageIndex, pageSize (optional): Page number and items per page, pageSize maximum 50.

• includeContent, contentMaxChars (optional): Content return switch and content length limit.

Returns: total and resultList.

5. information_topic

Function: Track specific topics such as policy, technology, brand, or hot events.

Main parameters:

• matchKeyword (required): Specific topic or event keyword.

• pubDateBegin, pubDateEnd (optional): Topic observation time range.

• pageIndex, pageSize (optional): Page number and items per page, pageSize maximum 50.

• includeContent, contentMaxChars (optional): Content return switch and content length limit.

Returns: total and resultList.

6. information_monitor

Function: Monitor enterprise information dynamics, with options for news event details or sentiment statistics views.

Main parameters:

• matchKeyword (required): Enterprise name, enterprise ID, registration number, or unified social credit code.

• view (optional): details news event details; statistics sentiment distribution and trends. Default statistics.

• keywordType (optional): Enterprise identifier type.

• pageIndex, pageSize (optional): Only used by the details view, pageSize maximum 50.

• sentimentLabel (optional): Sentiment filter only used by the details view.

Returns: The details view returns total and resultList; the statistics view returns sentiment distribution and trend business fields.

Use Cases

1. Enterprise identification: Confirm the full enterprise name and stable identifier via abbreviation or brand terms.

2. Information search: Aggregate global news titles, sources, times, and links by keyword.

3. Industry research: Track industry, track, and sector-related dynamics.

4. Topic tracking: Continuously observe policy, technology, brand, and hot events.

5. Enterprise public opinion: Query enterprise-related news and filter by sentiment category.

6. Trend monitoring: Analyze the sentiment distribution and change trends of enterprise information.

7. Brand reputation management: Identify reputation risks by combining news details and sentiment statistics.

Usage Notes

1. Abbreviation handling: When an enterprise abbreviation cannot be queried directly, first call information_enterprise_search.

2. Scenario selection: Use information_company_news or information_monitor for enterprise public opinion; use information_industry_news for industry research.

3. Pagination limits: pageIndex starts from 1, and pageSize must be between 1 and 50.

4. Content control: informationText is not returned by default; explicitly set includeContent=true when content is needed.

5. Content length: contentMaxChars must be between 100 and 5000.

6. Sentiment labels: 0 means negative, 1 means positive, 2 means neutral, 3 means unknown.

7. Valid empty results: When no keyword, time, or sentiment condition matches, total=0 and an empty resultList may be returned.

Example Usage Questions

information_enterprise_search (Enterprise keyword search)

1. Which enterprises correspond to "Xiaomi"?

2. Use "Gree" to find the exact enterprise name and enterprise ID.

information_search (Information full-text search)

1. Search for artificial intelligence-related news, looking only at titles and sources.

2. Search for new energy vehicle information and return two short content items.

information_company_news (Enterprise information query)

1. Query the latest enterprise news of Xiaomi Technology Co., Ltd.

2. View the negative public opinion of Gree Electric Appliances, Inc. of Zhuhai.

information_industry_news (Industry information search)

1. Search for the latest information in the low-altitude economy industry.

2. What are the recent industry developments in the humanoid robot track?

information_topic (Information topic tracking)

1. Continuously track the technology topic of large AI models.

2. Track events related to the consumer goods trade-in policy.

information_monitor (Enterprise information trend monitoring)

1. Monitor the public opinion sentiment distribution and trends of Xiaomi Technology Co., Ltd.

2. View the latest information event details of Beijing JD Century Trading Co., Ltd.

Testing and Verification

python -m py_compile server/mcp_server.py
python -m unittest discover -s tests -v

Unit tests use Mock HTTP responses and do not call the real HandaaS information API.