A recipe finder that works from what you already have. Enter the ingredients in your pantry, and ChefReserve searches a database of real recipes, ranks them by how many ingredients you already own, and lets you refine by cuisine, meal type, difficulty, spice level, and cook time.

1. You enter your pantry — type ingredients you have, optionally with quantities (e.g. 3 eggs, 6 oz beans)
2. Search — the backend queries Firestore for recipes that contain your ingredients, computes a match % for each, and returns the top 50
3. Refine — you can filter and rerank by cuisine, meal type, difficulty, spice level, and max cook time. Groq (Llama 3) reranks the filtered results
4. Recipe page — tap any recipe to see the full ingredient list (matched/missing highlighted), step-by-step instructions, and a link to the original source

Recipes are scraped from real cooking websites using the crawler in scripts/crawl_and_scrape.py. The crawler finds all recipe pages on a site by following links from the homepage, extracts structured data from schema.org/Recipe JSON-LD markup, and enriches each recipe with Groq to fill in cuisine, meal type, difficulty, and spice level.

Sites scraped:

• Epicurious
• Simply Recipes
• Serious Eats
• Budget Bytes

Each recipe is stored in Firestore with the following fields: name, source URL, cuisine, meal type, difficulty, spice level, cook time, servings, ingredients list, ingredient tokens (search index), and instructions.

chefreserve/
├── frontend/          React app (Vite)
│   └── src/
│       ├── pages/
│       │   ├── Home.jsx          Pantry input
│       │   ├── Results.jsx       Recipe list + preference modal
│       │   └── RecipePage.jsx    Full recipe detail
│       ├── components/
│       │   ├── RecipeCard.jsx    Recipe summary card
│       │   ├── PreferenceModal.jsx  Filter/refine UI
│       │   ├── PantryInput.jsx   Ingredient input
│       │   └── MatchBadge.jsx    Match % badge
│       └── lib/
│           ├── api.js            Backend API calls
│           └── firebase.js       Firestore client
│
├── backend/           Flask API
│   ├── routes/
│   │   ├── search.py    POST /api/search — pantry → candidate recipes
│   │   ├── rerank.py    POST /api/rerank — filter + Groq rerank
│   │   └── scraper.py   POST /api/scrape — single URL import
│   └── lib/
│       └── firebase_client.py
│
└── scripts/           Run locally
    ├── crawl_and_scrape.py   Bulk crawler + Groq enrichment
    └── requirements.txt

1. Create a project at console.firebase.google.com
2. Enable Firestore in Native mode
3. Go to Project Settings → Service Accounts → Generate new private key
4. Save the JSON file as backend/firebase-service-account.json
5. Copy your web config values to frontend/.env.local

1. Sign up at console.groq.com
2. Generate a free API key
3. Add GROQ_API_KEY=your_key to both backend/.env and scripts/.env

cd scripts
pip install -r requirements.txt
python crawl_and_scrape.py https://www.epicurious.com https://www.simplyrecipes.com

Tuning options (set in scripts/.env):

CRAWL_MAX_PAGES=500    # max pages per site (default 500)
CRAWL_MAX_DEPTH=3      # link depth from homepage (default 3)
CRAWL_DELAY=1.0        # seconds between page fetches (default 1.0)
ENRICH_DELAY=3.0       # seconds between Groq calls (default 3.0)

cd backend
pip install -r requirements.txt
cp .env.example .env   # fill in your keys
flask run

cd frontend
npm install
cp .env.local.example .env.local   # fill in your Firebase config + backend URL
npm run dev

backend/.env

FIREBASE_SERVICE_ACCOUNT=path/to/firebase-service-account.json
GROQ_API_KEY=your_groq_key

frontend/.env.local

REACT_APP_BACKEND_URL=http://localhost:5000
REACT_APP_FIREBASE_API_KEY=...
REACT_APP_FIREBASE_AUTH_DOMAIN=...
REACT_APP_FIREBASE_PROJECT_ID=...

scripts/.env

FIREBASE_SERVICE_ACCOUNT=path/to/firebase-service-account.json
GROQ_API_KEY=your_groq_key