FNewsCrawler is a FastAPI-based MCP service that aggregates real-time financial news and data from multiple sources. It implements MCP-compatible endpoints to expose tools for querying and processing financial information, including a general get_tool interface and a flexible AkShare integration endpoint. Clients can interact with the MCP server via HTTP to call specific tools (for example, stock data queries, news retrieval, and other AkShare-backed functions) and receive structured JSON results. The server also provides a web-based MCP tooling interface for enabling/disabling tools and monitoring usage, making it suitable for large models that need contextual financial data on demand.

To use it, first ensure the service is running and reachable at the configured address (default http://localhost:8480/mcp/mcp-server and API endpoints under /api/mcp). You can call a specific tool via the get endpoint, for example: GET http://localhost:8480/api/mcp/call_tool/{tool_name}?{params}, or use the AkShare bridge to invoke arbitrary AkShare functions through GET requests like: GET http://localhost:8480/api/mcp/call_akshare/{fun_name}?{params}. The documented examples show how to fetch industry funds flow, stock data, or news, and how to tailor results with parameters such as return_type, limit, filter_condition, and sort_by. The server also supports passing a list of URLs for batch news crawling and can return results in JSON or Markdown formats depending on the requested return_type.

Prerequisites:

• Python 3.8 or higher
• Redis server running locally or accessible
• Chrome/Chromium browser for headless crawling (as required by certain crawlers)

Installation steps:

1. Clone the repository: git clone https://github.com/noimank/FNewsCrawler.git cd FNewsCrawler

2. (Optional) Create and activate a virtual environment: python -m venv venv source venv/bin/activate # On Windows use: venv\Scripts\activate

3. Install Python dependencies (adjust if a requirements.txt is not present): pip install -r requirements.txt

   If there is no requirements file, install common dependencies such as FastAPI, uvicorn, httpx, redis, etc.

4. Configure environment:

   • Create .env and populate necessary variables (see the project’s .env.example): cp .env.example .env
   • Ensure Redis is reachable (host/port as required by your setup)

5. Run the MCP server:

   • Typical run (as used in the project docs): python main.py
   • Alternative Docker-based deployment can be used as described in the Docker deployment section of the docs.

6. Verify the server is up:

   • Open http://localhost:8480/monitor or http://localhost:8480/mcp to check status
   • Call an API endpoint e.g. http://localhost:8480/api/mcp/call_tool/get_industry_stock_funds_flow_tool?industry_name=银行 to test functionality

Notes and tips:

• The MCP server supports both SSE and HTTP-based access; you can switch modes using the MCP_SERVER_TYPE environment variable.
• If you run locally, ensure Redis is started before launching the MCP service, as many features depend on Redis for state persistence.
• Some tools may require additional configuration or network access to external data sources (AkShare, news sources, etc.).
• When using the AkShare endpoints, parameters should avoid enclosing quotes unless required by the API (see the API docs in the README for examples).
• For production, consider running behind a reverse proxy and enabling health checks. Monitor endpoints at /monitor and /logs for troubleshooting.
• If you switch to the Docker deployment, follow the Docker deployment guide found in the project’s docs or docker/README.md for complete options and environment variable configurations.