This MCP server is a demonstration that streams the top Weibo hot searches using the fastmcp framework. It exposes three running scripts for different deployment scenarios: Stdio, SSE, and Streamable HTTP. Stdio runs locally via standard input/output, SSE provides a server-sent events stream for remote connections, and Streamable HTTP serves a long-lived HTTP response suitable for custom frontends or other MCP clients. After starting any of the scripts, you can call hot_search(n=20) to fetch the top n hot searches. The example uses the in-repo resou_stdio.py script for local development and can be adapted to deploy the same logic over different transport modes as needed.

Prerequisites:

• Conda (or Miniconda) available on your system
• Python 3.11

Step-by-step installation:

1. Clone the repository git clone https://github.com/bossdong955/mcp-server-example.git cd weibo-mcp-server

2. Create and activate the Python environment conda create -n mcp-server-example python=3.11 conda activate mcp-server-example

3. Install dependencies pip install -r requirements.txt

4. Run the desired script (examples shown below):

   Stdio (local)

   python resou_stdio.py

   SSE (remote streaming)

   python resou_sse.py --host 0.0.0.0 --port 8006

   Streamable HTTP

   python resou_streamable_http.py --host 0.0.0.0 --port 8005

Note: If you previously installed a conflicting mcp package, uninstall it first with: pip uninstall mcp

Tips and notes:

• For encoding issues, set PYTHONUTF8=1 in your environment or just in the command preceding the Python invocation.
• If you encounter dependency conflicts, ensure you install within a clean Conda environment as shown and remove any globally installed old mcp packages.
• The default hot_search timeout is 10 seconds when using httpx; adjust the timeout in the code if you need longer waits.
• In SSE mode, you can specify custom Authorization headers in the client if your server requires authentication.
• The three modes share the same underlying logic; choose the transport that best fits your deployment scenario (local development vs remote streaming vs HTTP streaming).