FridgeChef AI helps people answer “What can I cook with what I already have?” and make better use of their ingredients.

Upload a fridge photo, review the ingredients identified by Gemma, choose dietary preferences, and generate personalized recipes with cooking instructions and serving sizes.

Features:

• Photo-based ingredient recognition and recipe generation using Gemma through Ollama.
• Vegetarian, vegan, gluten-free, and allergy/exclusion preferences, with ingredient conflict checks.
• Estimated calories, protein, carbohydrates, fat, and fiber calculated from a local USDA nutrition dataset. Incomplete estimates are clearly marked.
• Locally stored serving-suggestion photos with illustration fallbacks.
• Saved recipes, recipe copying, sorting, and light/dark themes.

Built with Python, Flask, and vanilla HTML, CSS, and JavaScript. AI inference runs locally, without a paid API or cloud account. After initial setup and model download, the application can run offline.

Nutrition values are estimates. Allergy checks are an additional aid and do not guarantee the absence of allergens or cross-contact.