接入你的智能体
從發布者安裝。不是 PlugMCP 端點。
遠端 URL 還是本機指令? 兩種安裝方式的差異
A small experimental project exploring two things at once:
CIQUAL is the French food composition database published by ANSES. It contains nutritional data (macronutrients, vitamins, minerals, energy, etc.) for thousands of food items, per 100g. This project bundles the 2025 release of CIQUAL as XML files and parses them into an in-memory database.
The project is a Maven multi-module application with two Spring Boot services:
CiqualMcp/
├── CiqualMcpServer/ ← MCP server exposing food search tools (port 8080)
└── CiqualMcpClient/ ← MCP client + web chat interface (port 8081)
CiqualMcpServer)searchFood tool via MCP using spring-ai-starter-mcp-server-webmvc.CiqualMcpClient)spring-ai-starter-mcp-client.gpt-4o-mini) through Spring AI's ChatClient.POST /api/chat) and a simple web UI for chatting with a nutritional assistant.searchFood tool on the MCP server to fetch real CIQUAL data and incorporates it into its answer.User (browser)
└─► ChatController (port 8081)
└─► Spring AI ChatClient → OpenAI gpt-4o-mini
└─► [tool call] MCP Client
└─► MCP Server (port 8080)
└─► FoodSearchTools → EnrichedCiqualDatabase
└─► SemanticFoodSearch (fuzzy + synonym matching)
└─► CIQUAL XML data
cd CiqualMcpServer
mvn spring-boot:run
The server starts on port 8080. It will take a moment to parse the CIQUAL XML files (the composition file is ~67 MB).
cd CiqualMcpClient
OPENAI_API_KEY=sk-... mvn spring-boot:run
The client starts on port 8081.
Open http://localhost:8081 and ask something like:
This is intentionally small and exploratory. The main goals were:
Spring AI MCP server/client for querying the French CIQUAL food database via natural language
工具
尚無工具清單。遠端伺服器會在首次健康檢查後取得工具清單。
常見問題
在本頁複製適用於 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 設定。它使用發布者的安裝方式,而不是 PlugMCP 代理伺服器。
這筆記錄未列出必要的環境變數。
CiqualMcp 透過 stdio 在本機執行。
上次檢查時間:2026-10-28。 Spring AI MCP server/client for querying the French CIQUAL food database via natural language