The problem with generic recipe recommendations
Most recipe recommendation systems work like a popularity contest. They surface what's trending, what users broadly rate highly, or what fits a loose category search. None of that tells you whether a recipe is right for you — your diet, your cuisine preferences, your existing collection, or what you're actually missing nutritionally.
The result is a feed of interesting recipes that rarely feels personalised. You save a few, never make them, and the cycle repeats.
How Mealible's AI starts from your collection
Mealible's AI begins by analysing the recipes you've already saved. It reads the nutrient profiles, identifies which foods you cook regularly, and maps out your nutritional coverage — which nutrients are well-represented in your typical week, and which are consistently missing.
This is the key difference: instead of recommending popular recipes, the AI recommends recipes that fill specific gaps in what you actually cook.
Gap detection: more nuanced than you might expect
A simple gap detector would just find your lowest-scoring nutrients and recommend foods rich in those. But Mealible's approach is more sophisticated. It distinguishes between two types of gaps: quantity gaps (not enough fiber-rich recipes) and diversity gaps (plenty of protein recipes, but they all use the same source — mostly chicken).
Source diversity matters nutritionally. Eating chicken every day provides protein but misses the iron of red meat, the omega-3s of fish, and the fiber and micronutrients that come with plant proteins. Mealible prioritises diversity gaps even when quantity looks fine.
Dietary constraints respected automatically
The AI infers your dietary preferences from your saved recipes. If most of your saved recipes are vegetarian, it won't recommend salmon to fill a protein gap — it will look for lentil, paneer, or egg-based recipes instead. No explicit settings to configure; the system reads your habits.
This matters particularly in the UAE context, where halal requirements and diverse culinary preferences mean a generic recommendation system would frequently suggest inappropriate options.
10 picks, weekly, grouped by gap type
Each week, the AI produces 10 recommendations grouped by the gap they address — protein variety, fiber, vitamins, or new cuisines to explore. Each recommendation comes with a plain-English explanation of exactly why it was chosen for your collection.
You can save a pick (it goes straight into your recipe library) or dismiss it (the AI notes this preference for future recommendations). Over time, the picks become more aligned with what you actually want to cook.
The result: a more balanced recipe library, automatically
The goal isn't to show you interesting recipes — it's to gradually shift your recipe collection towards better nutritional coverage. More variety in your saved recipes means more balanced meal plans, which means a more nutritious week — without ever having to manually track a nutrient or count a calorie.