MacroBud

Tracking macros gets tiring, and when we eventually all give up, we just start estimating in our heads. Macrobud estimates for you, giving more accurate figures so you actually hit your goals.

Prerequisites

• Node.js (v18+)
• Python (3.10+)

No database install needed -- the app uses SQLite which is built into Python.

Setup

1. Backend

cd backend
pip install -r requirements.txt

Add your OpenAI API key to backend/.env:

OPENAI_API_KEY=sk-your-key-here

Start the server:

uvicorn main:app --reload

The API runs at http://localhost:8000.

2. Frontend

cd frontend
npm install
npm run dev

The app runs at http://localhost:5173.

Usage

1. Open http://localhost:5173 in your browser.
2. Type one or more foods into the input (e.g. "1 slice of rye bread, 2 eggs, a glass of milk").
3. The app sends your input to OpenAI, which splits multi-food entries and estimates macros for each item.
4. If anything is vague (e.g. "chicken breast" without a quantity), you'll get a prompt to clarify.
5. Your running totals update as you add or remove items.

Tech Stack

Layer	Technology
Frontend	React + Vite
Backend	Python + FastAPI
Database	SQLite
AI	OpenAI gpt-4o-mini