McDonald's Nutrition Diet Expert

McDonald's Nutrition Diet Expert

Picks McDonald's meals that fit your daily calorie and sodium targets, then helps you order and log them, for people managing their weight or building muscle.

Made by clawgeek on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.

About

What it is: This is a diet tool that picks McDonald's meals to fit your body and your goal. It plots menu items on a matrix of nutrition density against sodium load, then recommends meal combinations. Because it sits on McDonald's official MCP server, it also covers store menus, coupons, ordering and a health log.

Who it suits: It suits people who want to eat at McDonald's while managing weight, muscle or sodium. Developers can also use it as a working example of MCP in a service.

Industry | Who uses it | How they use it | Example
Weight management | People cutting fat | Get a meal sized to a daily calorie target, with a low-sodium option | A 626 kcal target for one meal, with a 555 kcal plan
Fitness | People building muscle | Find high-protein meals and see which ones are also high in sodium | A lunch after leg day without a sodium spike
Health management | People watching sodium | Rank combos so that high-sodium ones score lower | A combo over a 2000 mg sodium limit gets flagged
Software | Developers learning MCP | Reuse the source, architecture notes and API docs in other MCP projects | Five documented API traps

Anyone who does not eat at McDonald's can skip it.

Where to use it: It runs as a web app on your own machine, as a REST API and as a WorkBuddy skill.

• Web app: a browser interface with the matrix, meal plans, store menus, a coupon wallet, a cart and a health log
• WorkBuddy: delivered as a skill that you call in plain language inside a WorkBuddy chat
• REST API: a FastAPI server with 14 endpoints you can call yourself
• Your own machine: the server runs locally from the project folder

Use case | Who | What they get back
Choosing a meal at the ordering kiosk | Everyday users | The one-click solve gives three meal plans, with a quadrant badge on each item
Reviewing the calorie gap | People who log meals | A weight trend, calorie gap and quadrant mix
Coupons and ordering | Anyone ordering at a store | Held coupons in a wallet, a live-priced cart and order submission
Learning MCP | Developers | Full source code, architecture notes and API docs

How to set it up: You need Python and a token for McDonald's MCP server before the app can start.

• Python 3.10+
• An MCP token, which goes in a .env file
• git and pip

1. Clone the repository from GitHub.
2. Open the skill folder inside it.
3. Copy the example environment file to .env and add your MCP token.
4. Install the Python requirements.

These commands cover steps 2 to 4:

cd mcd-nutrition-expert/skills/mcd-nutrition-expert
cp .env.example .env
pip install -r requirements.txt

How to use it: Start the server and open the web app in your browser. It listens on port 8899 on your own machine.

python scripts/server.py

In the app you build a profile, read the matrix, pick one of three plans, choose a store, claim coupons, place the order and log the meal afterwards. In WorkBuddy you can ask for a meal by the SP matrix, ask how to order today, or ask whether your calorie gap covers a plan.

Tool | What it does
Profile | Turns your height, weight and age into a daily target and a target point on the matrix
SP matrix | Plots menu items in four quadrants, with a gold cross at your target
Recommendation | Gives three meal plans, with a fit percentage for each item
Store ordering | Shows a store's menu with a quadrant badge and fit on each item
Coupons | Shows the coupons you hold and the ones you can claim
Cart and order | Prices the cart live and submits the order
Health log | Records what you ate and shows weight trend, calorie gap and quadrant mix

Example: Enter a profile of a 28-year-old woman, 165 cm and 58 kg, moderately active and cutting fat. The tool sets a target of 1625 kcal a day and 552 kcal a meal. Its first plan is 637 kcal with 44 g of protein and 1878 mg of sodium, and it warns that this reaches 94% of the limit. The second plan is 580 kcal with 1049 mg of sodium, about half, and 11 g less protein.

For newcomers and experts: If you are new to this, start with the web app. Enter your height, weight and age, and it shows a daily calorie target and a scatter of menu items with a gold cross at your target. Then open the three meal plans to see what each one trades away.

Under the hood, the matrix is a strategy chart reused for food. One axis is nutrition density, protein plus carbs per 100 kcal, and the other is sodium per 100 kcal. Your daily target comes from the Mifflin-St Jeor equation with FAO/WHO activity factors, and one meal gets 34% of it. An item's fit score falls as its distance from your target point grows, so the same item fits different people differently. Plans come from a constraint solver over meal roles, and a plan that breaks the sodium cap is flagged in its warnings instead of hidden.

Profile gives a daily and a per-meal target
Menu items are placed on the SP matrix
A constraint solver builds three meal plans
You pick a store, claim coupons and order
The meal log tracks weight and calorie gap

Good to know:

• Data: nutrition data for 160 items, a store menu of 98 items, and 88% of menu items placed on the matrix
• Official tools: it uses 35 tools from McDonald's official MCP server
• Storage: daily records and weight trend are kept in SQLite
• Front end: the web interface has no third-party dependencies
• Language: the README and the sample meals are in Chinese
• License: MIT

Title: Menu items per quadrant
Main zone: 47 items
Craving zone: 46 items
Filler zone: 53 items
Double-edged zone: 10 items

On GitHub, 7 November 2026

• Stars: 104
• Forks: 1
• Open issues and pull requests: 0
• Language: Python
• License: MIT
• Topics: canvas, diet, fastapi, fitness, hackathon, health, mcd-mcp, mcdonalds, mcp, mcp-server, meal-planner, model-context-protocol

Links

• Code: https://github.com/clawgeek/mcd-nutrition-expert
• Website: https://github.com/ChinaOpenClaw/mcd-nutrition-expert#readme
• Releases: https://github.com/clawgeek/mcd-nutrition-expert/releases
• Issues: https://github.com/clawgeek/mcd-nutrition-expert/issues
• Star history: https://www.star-history.com/#clawgeek/mcd-nutrition-expert&Date

Get the code

git clone https://github.com/clawgeek/mcd-nutrition-expert.git

Then follow the setup in its README. Nothing in this kit runs by itself.