v1.0.0 • Stable

Make food data talk to your LLM.

Edamam MCP is a unified, LLM-friendly interface on top of the Edamam Food Database API. Use it to search foods, fetch nutrition, and analyze food images via REST, JSON-RPC, or a local Inspector UI.

Get started with the API Get API Keys Try the Inspector UI

REST + JSON-RPC + MCP schema

LLM-ready function definitions

Backed by Edamam Food Database

MCP in action

REST

JSON-RPC

# REST: get nutrition for 100g banana
POST /v1/ai/query
Content-Type: application/json

{
  "intent": "get_food_nutrition",
  "parameters": {
    "query": "banana",
    "quantity": 100
  }
}

Or use JSON-RPC via POST /v1/rpc for tools/agents.

Capabilities

What Edamam MCP does for you

A compact, opinionated interface covering the most common food and nutrition use cases, designed from day one for LLMs and agents.

🍌

Search foods

Search by food name or barcode (UPC/EAN/PLU). MCP normalizes the result into a consistent structure for your client or LLM.

🥗

Get full nutrition

Fetch total nutrients for a given food and quantity. Ideal for nutrition apps, meal planners and health platforms.

📸

Image recognition

Analyze food images and get ingredients, serving weight and nutrient breakdown via Edamam’s food-vision capabilities.

🔁

Dual transport

Use classic REST (`/v1/ai/query`) or JSON-RPC (`/v1/rpc`) to fit your orchestration layer, agents or tool framework.

🧠

MCP schema

`/v1/mcp/schema` exposes a ready-to-use system prompt + function definitions that plug directly into your LLM.

🧪

Inspector UI

Visual tooling to test MCP calls, tweak arguments, save presets, replay requests and export as cURL.

Why an MCP?

Food APIs are REST. Your stack is LLM.

Bridging Edamam’s Food Database API into an AI / agent stack usually means glue code, documentation hunting, and repetitive boilerplate. Edamam MCP removes that friction.

Raw REST

Scattered endpoints

Different REST endpoints, parameters and response shapes for search, nutrition and image analysis. Great for humans, noisy for LLMs.

Manual glue

Custom glue code everywhere

You write and maintain the same “tiny adapter” logic over and over for each app, environment, or experiment.

With Edamam MCP

One clean interface

A single, LLM-friendly surface: tools, intents, schema, REST and JSON-RPC transport, plus a local Inspector UI for instant testing.

How it works

Three ways to talk to Edamam MCP

Pick the transport that matches your stack: developer-friendly REST, tool-friendly JSON-RPC, or LLM-native MCP schema.

1. REST / HTTP

Use POST /v1/ai/query with an intent and parameters. Perfect for backends and internal services.

2. JSON-RPC / MCP tools

Use POST /v1/rpc with tools/list and tools/call. Ideal for AI runtimes, agents, and tool frameworks.

3. MCP schema for LLMs

Call GET /v1/mcp/schema at startup to auto-discover the system prompt and tool definitions. Register them as model tools and let the LLM decide what to call.

REST – get nutrition POST /v1/ai/query

curl -X POST http://localhost:8000/v1/ai/query \
  -H "Content-Type: application/json" \
  -d '{
    "intent": "get_food_nutrition",
    "parameters": {
      "query": "almonds",
      "quantity": 50
    }
  }'

JSON-RPC – search food POST /v1/rpc

curl -X POST http://localhost:8000/v1/rpc \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "search_food",
      "arguments": {
        "query": "banana",
        "limit": 5
      }
    },
    "id": 1
  }'

MCP schema – for your LLM GET /v1/mcp/schema

curl http://localhost:8000/v1/mcp/schema

{
  "mcp_name": "mcp-edamam",
  "mcp_version": "1.0.0",
  "description": "...",
  "functions": [ ... ],
  "system_prompt": "You are connected to Edamam MCP ..."
}

Inspector UI

A local lab for your MCP

Edamam MCP ships with a small React-based Inspector. It connects to your local MCP instance and lets you visually test tools, debug requests and share presets with your team.

What you can do

• Switch between REST and JSON-RPC transports
• Pick a tool, fill parameters, and execute
• See the raw request / response JSON
• Save presets and replay them later
• Export calls as ready-to-run cURL commands
• Use auto-retry for flaky connections

Run the Inspector UI Local dev

cd inspector-ui
npm install
npm run dev

# Open in your browser:
# http://localhost:5173

Point the Inspector at your MCP backend (e.g. http://localhost:8000) and start exploring the schema, tools and responses without writing a single line of glue code.

Docs & API

Documentation you can actually use

Whether you’re integrating Edamam MCP into a production stack or experimenting with a local agent, you get clear endpoints, examples and wiring diagrams.

Endpoints at a glance

GET /v1/mcp/schema – MCP metadata and tools
POST /v1/ai/query – unified REST execution
POST /v1/rpc – JSON-RPC 2.0 tools endpoint
GET /docs – Swagger / OpenAPI UI

Full reference, examples and LLM integration patterns are available in the documentation bundle.

Edamam Food Database API Edamam Data Licensing

Docs bundle

01 – Overview
02 – API reference
03 – Examples
04 – LLM integration
05 – Inspector UI guide

Stored as markdown alongside the MCP codebase for quick navigation and versioning.

View docs

Download

Get Edamam MCP v1.0.0

Clone the repository, download a tagged release, or pull it into your internal infrastructure. Edamam MCP is designed to run close to your LLM stack.

Releases & archives

v1.0.0 – first stable release (REST + JSON-RPC + Inspector)
View changelog
Browse on GitHub

Older versions and pre-release builds can be accessed via the GitHub Releases section.

Download MCP

Get the latest stable version as a source archive or clone it directly from GitHub.

Download v1.0.0 Get API Keys

Requires Python 3.10+, FastAPI, and a valid Edamam API key / app ID.