pleytj/mcp-food-analyzer

🧠 MCP Server – Profile Analyzer Tool

This project is a cloud-hosted, MCP-compatible microservice built with FastAPI. It exposes an endpoint that analyzes food products against a user's dietary profile — perfect for use with tools like n8n, chatbots, or health apps.

🚀 What This Does

This service exposes a single endpoint:

POST /mcp/profile_analyzer

It receives a structured JSON body containing:

• A user's dietary profile (allergies, diet, health goals, dislikes)
• A list of ingredient IDs
• Nutrition values per 100g

It returns:

• A score (0–100)
• flags indicating issues (e.g. "contains_peanut", "high_sugar")
• A summary and explanation
• Smart suggestions and product alternatives

🛠️ Tools Used

🔗 Live URLs

🔄 Example Request

POST /mcp/profile_analyzer
Content-Type: application/json

{
  "user_profile": {
    "allergies": ["peanut"],
    "diet": ["vegan"],
    "dislikes": [],
    "health_goals": ["reduce sugar"]
  },
  "ingredients": [
    { "id": "peanut" },
    { "id": "cocoa" }
  ],
  "nutrition_per_100g": {
    "sugars": 18.0
  }
}

🧠 Why Use MCP?

Using an MCP (Model Control Protocol) structure allows us to:

• Modularize AI tools as callable APIs
• Reuse the same logic across workflows and platforms
• Swap models or prompts without frontend changes
• Scale into more advanced tools later (like GPT-4 Vision or CrewAI agents)

📁 Project Structure

/main.py                   # Starts the FastAPI app
/tools/
  └── profile_analyzer.py  # Core analysis logic
/requirements.txt          # Python dependencies

📝 Next Steps

• Add more tools (e.g., /mcp/label_reader)
• Connect this API to n8n workflows
• Add OpenAI integration for GPT-based tools
• Secure endpoint with API key if needed