Edamam Remote Food MCP

Food APIs are built for apps. MCP makes them useful inside agent workflows.

Edamam Food MCP turns the Food Database API into a small set of LLM-friendly tools. The agent keeps the conversation natural, while the server handles food parsing, quantity resolution, UPC lookup, nutrition calculation, and image-based dish analysis.

01

One endpoint for agents

Use a single MCP endpoint for initialize, tools/list, and tools/call. No custom adapter layer for every assistant or automation tool.

02

Nutrition from plain language

Send phrases like "cheddar cheese", "1 cup cooked rice", "200g chicken breast", or a UPC/EAN barcode.

03

Food image analysis

Provide a public image URL or base64 data URI and return detected food plus nutrition estimates from Edamam's Vision API.

The agent asks naturally. The MCP server returns structured nutrition.

User

I had a chicken enchilada and a French onion soup with a slice of bread for lunch. How many calories in this

Agent

Based on the nutrition data provided (1 whole serving each):

Calories (lunch total)

• Chicken enchilada (1 whole): 503.7 kcal
• French onion soup (1 whole): 395.6 kcal
• Bread (1 slice): 79.5 kcal

Total calories: 978.7 kcal (approx 979 kcal)

get_food_nutrition({
  "queries": ["1 chicken enchilada", "1 French onion soup", "1 slice of bread"]
})

User

Give me macro breakdown and what diets is this good for?

Agent

Macro breakdown (whole lunch)

• Protein: 53.0 g
• Carbs: 81.8 g
• Fat: 47.8 g

Diet fit: Best for a general omnivorous/balanced diet. This meal is not a strong fit for keto/low-carb, low-fat, low-sodium, vegan, vegetarian, gluten-free, or dairy-free eating patterns (includes wheat/gluten and milk cautions in the returned items).

MCP Agent Nutrition Assistant Demo

Look up nutrition facts for one or more foods described in plain text. Identify food in a image and return a full nutritional breakdown.

Connect your MCP client in minutes.

Add the server URL to your client and pass Edamam credentials as a bearer token. Discovery can connect first; actual tool calls require an Edamam Food Database API app_id and app_key.

{
  "mcpServers": {
    "edamam-food": {
      "url": "https://mcp.edamam.com/mcp/food",
      "headers": {
        "Authorization": "Bearer <app_id>:<app_key>"
      }
    }
  }
}

{
  "servers": {
    "edamam-food": {
      "type": "http",
      "url": "https://mcp.edamam.com/mcp/food",
      "headers": {
        "Authorization": "Bearer <app_id>:<app_key>"
      }
    }
  }
}

claude mcp add --transport http edamam-food \
  https://mcp.edamam.com/mcp/food \
  --header "Authorization: Bearer <app_id>:<app_key>"

{
  "mcpServers": {
    "edamam-food": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.edamam.com/mcp/food",
        "--header",
        "Authorization: Bearer <app_id>:<app_key>"
      ]
    }
  }
}

Two production tools, tuned for food questions.

The FastMCP server keeps the agent surface intentionally compact. Search, parsing, UPC recognition, and nutrient calculation are handled inside the nutrition tool; photos use a dedicated Vision tool.

tool

get_food_nutrition

Look up calories, macros, diet labels, health labels, cautions, total nutrients, daily values, and resolved ingredient data for one or more foods.

Argument	Type	Required	Description
queries	list[string]	Yes	One phrase per food item, such as "1 cup rice", "200g chicken breast", or "049000042566".

Example JSON-RPC call

{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "get_food_nutrition",
    "arguments": {
      "queries": ["1 croissant", "1 egg"]
    }
  },
  "id": 3
}

tool

analyze_food_image

Identify food in a photograph and return the detected item, quantity, measure, ingredient lines, calories, labels, cautions, and full nutrient breakdown.

Argument	Type	Required	Description
image	string	Yes	A public http / https URL or data:image/... base64 URI.

Example JSON-RPC call

{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "analyze_food_image",
    "arguments": {
      "image": "https://example.com/plate.jpg"
    }
  },
  "id": 4
}

Use your Edamam credentials without storing them on the MCP server.

Tool calls use the Food Database API credentials supplied by the caller. Send a bearer token with the app_id and app_key joined by a colon.

Authorization: Bearer <app_id>:<app_key>

For self-hosting: set EDAMAM_APP_ID and EDAMAM_APP_KEY as fallback credentials, or require every caller to pass the Authorization header.

Everything needed to test, self-host, and integrate.

The public endpoint is for immediate MCP clients. The GitHub repository contains the FastMCP server, shared auth and logging code, pytest coverage, deployment notes, and examples for direct JSON-RPC calls.

Get API keys Food Database API docs MCP Streamable HTTP spec

1

Initialize

Client sends initialize to negotiate capabilities and receives an MCP-Session-Id.

2

Discover tools

Client calls tools/list and receives schemas for get_food_nutrition and analyze_food_image.

3

Call tools

Client sends tools/call with the session id and Edamam bearer credentials.

Mapped to the Food Database API paths agents need most.

Food Database capability	MCP mapping	Status
Parser search by food name or UPC	Internal step in get_food_nutrition	Available
Nutrients endpoint	get_food_nutrition	Available
Nutrients from image	analyze_food_image	Available
Autocomplete and browse filters	Not exposed as MCP tools	Out of scope

Included with Edamam Food Database API plans.

Edamam lists Food MCP as included with Food Database API plans. Choose the plan by API volume, Vision usage, throttling, caching rights, and support needs.

EnterpriseBasic Vision

$14 / month

• 100,000 calls/month
• 50 food and nutrition calls/min
• 500 Vision requests/month
• 30 day trial period

EnterpriseCore

$69 / month

• 750,000 calls/month
• 100 food and nutrition calls/min
• Vision pay as you go
• Email and phone support

EnterprisePlus

$299 / month

• 5,000,000 calls/month
• 300 food and nutrition calls/min
• 10,000 Vision requests/month
• Email and phone support

Plan details change over time. Confirm current prices, quotas, caching rights, and terms on the Edamam Food Database API page before launch.

Read-only food data, with API-key responsibility at the client edge.

Keep app keys private

Never commit app_key values to source control or paste them into shared prompts. Treat them like passwords.

Human initiated usage

Edamam API terms restrict automated collection, scraping, or building a copy of returned data.

Attribution required

When showing Edamam nutrition data to end users, display the required Edamam attribution and link.

Informational output

Nutrition and Vision results should be presented as estimates for product use, not as medical or dietary advice.

Common setup questions.

Do I need Edamam API keys before connecting an MCP client?

You can connect to the public Food MCP endpoint and discover available tools first. To run actual tool calls, you need an Edamam Food Database API app_id and app_key sent as a Bearer token.

How do I subscribe to a Food MCP plan?

Choose a Food Database API plan from the pricing section or the signup modal, create your API account, and then log in to complete the next billing step for API access.

Does the server store my Edamam credentials?

No. Public-client calls pass credentials in the Authorization header. Self-hosted deployments may set environment variables as fallback credentials.

Why is there no standalone search_food tool?

The current FastMCP implementation resolves food search internally inside get_food_nutrition. A separate search-only tool can be added later if user workflows need browse-style results.