This MCP server provides access to Baidu Cloud Content Safety for large models via the MCP (Model Context Protocol). It exposes an input_analyze tool that analyzes user input text for safety concerns using Baidu's content safety capabilities. To run it, install Python 3.10+, ensure uv (the MCP runner) is installed, clone the repository, and start the server using uv. You can configure the server with your Baidu Cloud credentials via environment variables or a .env file. The server is designed to be consumed by MCP-enabled applications such as Cursor or any MCP-compliant client, enabling seamless content safety checks as part of your model workflow.

Prerequisites:

• Python 3.10 or newer
• uv (MCP runner) installed: pip install uv

Installation steps:

1. Clone the repository git clone https://github.com/liangjunyu2010/mcp_server_safe_content_check.git cd mcp_server_safe_content_check

2. Install dependencies (if any are listed in a requirements file; otherwise uv is the primary runtime dependency)

   If a virtual environment is preferred:

   python -m venv venv source venv/bin/activate # on Windows: venv\Scripts\activate pip install -r requirements.txt # if available

3. Prepare credentials

   • Set environment variables when starting the server or via a .env file: BAIDU_CLOUD_ACCESS_KEY_ID=YOUR_ACCESS_KEY_ID BAIDU_CLOUD_SECRET_ACCESS_KEY=YOUR_SECRET_KEY

4. Run the server

   • Using uv directly as recommended in the repository: uv run src/mcp_server_safe_content_check/server.py

5. (Optional) Run with explicit credentials on the command line: uv run src/mcp_server_safe_content_check/server.py --BAIDU_CLOUD_ACCESS_KEY_ID YOUR_ACCESS_KEY_ID --BAIDU_CLOUD_SECRET_ACCESS_KEY YOUR_SECRET_KEY

Notes:

• If you prefer Cursor or another MCP client, configure the MCP server entry with the appropriate command/args as shown in the README. Ensure the environment variables are available to the runtime.
• You can also place credentials in a local .env file and rely on uv's environment-loading behavior if supported in your setup.

Tips and common issues:

• Ensure Python 3.10+ is installed and accessible in your PATH.
• If using a virtual environment, activate it before running uv.
• When configuring via Cursor, use the provided mcp.json snippet to register the MCP server with command 'uv' and the given arguments.
• If you encounter authentication errors, verify that BAIDU_CLOUD_ACCESS_KEY_ID and BAIDU_CLOUD_SECRET_ACCESS_KEY are correct and have the necessary permissions.
• The server expects the Baidu Cloud credentials to be available as environment variables; you can also export them in your shell or place them in a .env file read by uv/CICD pipelines.
• The input_analyze tool accepts a single parameter: text (the content to analyze). You can extend or customize the MCP config to expose additional tools if supported by the server.
• When integrating with Cursor, you may need to restart Cursor after adding the new MCP server configuration to reload the mcp.json.