Know exactly which menu items make money
AI analyzes your menu photos and pricing to surface your most and least profitable dishes in minutes—not hours.
$pip install plate-profit Snapshot any menu item
Upload a photo of your menu or enter item names. GPT-4V extracts dish names, ingredients, and portions. USDA price data estimates ingredient costs.
- Works with printed, digital, or chalkboard menus
- No manual data entry required
- Cross-references with live USDA market prices
```python
from plate_profit import MenuAnalyzer
analyzer = MenuAnalyzer()
result = analyzer.analyze_photo("menu.jpg")
print(result.profitable_items)
# [{'dish': 'Margherita Pizza', 'margin': 0.72}, ...]
```Calculate plate costs instantly
Break down each dish into ingredient-level costs. See food cost percentage, gross margin, and contribution margin side-by-side.
- Ingredient-level cost modeling
- Portion-size awareness
- Waste factor estimation built in
```python
dish = analyzer.calculate_plate_cost(
"Chicken Tikka Masala",
ingredients=["chicken_breast: 200g", "tomato: 150g"],
selling_price=18.95
)
print(f"Food cost: {dish.food_cost_pct:.1%}")
print(f"Gross margin: ${dish.gross_margin:.2f}")
# Food cost: 28.4%
# Gross margin: $13.54
```Engineer your menu mix
Rank every item by profitability. Identify overpriced losers and underpriced winners. Get AI suggestions to optimize your menu board.
- Profitability ranking across all items
- Price elasticity recommendations
- Export to CSV or POS sync
```python
menu = analyzer.analyze_full_menu()
# Sort by margin, show bottom 5
losers = menu.sorted_by_margin().bottom(5)
for dish in losers:
print(f"{dish.name}: {dish.margin:.0%}")
# Get AI suggestions
suggestions = analyzer.get_pricing_recommendations(losers)
```Common questions
How accurate are the cost estimates?
Accuracy depends on photo quality and ingredient clarity. GPT-4V extracts dish components from photos, then cross-references with USDA wholesale price data. For complex recipes with 10+ ingredients, accuracy is ~85% vs manual calculation. We recommend validating top-revenue items manually during your first week.
What data does Plate Profit access?
You upload menu photos to our API, which sends them to OpenAI for vision analysis. We do not store your photos. Cost data comes from public USDA databases—we do not access your POS or financial systems unless you explicitly connect them via integration.
Can I run this offline or on-premises?
Currently Plate Profit runs as a cloud API. For operators with data-privacy requirements, we offer a self-hosted option using local LLM inference. Contact us for pricing on the on-premises tier.
Does this work for multi-location operators?
Yes. You can batch-upload menus from multiple locations and compare profitability side-by-side. Each location is treated as an independent menu in our system.
What does the $500 include?
The starter package includes 500 menu scans, full cost analysis for up to 200 dishes, and one export to CSV or POS-compatible format. Monthly subscriptions for higher volume start at $99/month.
Stop guessing which dishes profit
Join 500+ operators using Plate Profit to engineer menus that actually make money.