The Bocha AI Web Search MCP Server exposes a web search capability built around the Bocha AI service. It provides a tool named bocha_web_search that retrieves content from Chinese internet sources, with optional detailed summaries and image results. The output can be delivered in two formats: Markdown for human readability, including links and image previews, or JSON for programmatic processing. To use it, issue a search query through the bocha_web_search tool with parameters such as query, freshness, summary, count, and raw_json to control the output format. The server returns structured results that include web content, summaries, and related images as applicable. If you prefer raw programmatic data, set raw_json to true to receive a JSON structure instead of Markdown.

Prerequisites:

• Node.js and npm installed on your machine
• Access to Bocha AI API and a valid API key

Step-by-step installation:

1. Clone the MCP server repository: git clone git@github.com:yoko19191/bocha-ai-mcp-server.git
2. Navigate to the project directory: cd bocha-ai-mcp-server
3. Copy the sample environment file and configure your API key: cp .env.example .env

   Edit .env and set BOCHA_API_KEY to your Bocha AI API key

4. Install dependencies: npm install
5. Build the server (if a build step exists): npm run build
6. Run the MCP server via Node.js (as described in the Configuration section): node /ABSOLUTE/PATH/bocha-ai-mcp-server/build/index.js

Notes:

• Ensure your BOCHA_API_KEY is set in the environment where the server runs.
• The example configuration uses an absolute path to the built server file; adjust as needed for your deployment environment.

Tips and considerations:

• Ensure the BOCHA_API_KEY is kept secret and not exposed in client-side code.
• The server supports both Markdown and JSON outputs; use the raw_json flag to switch formats.
• If you encounter issues with API rate limits or connectivity, verify network access to Bocha AI endpoints and validate your API key.
• When deploying in production, consider setting up environment-based configuration management and logging to monitorBocha API usage and response times.
• The mcp.json example uses a single server named bocha-ai-search; you can add additional MCP servers by extending the mcpServers map.