How DinDin works
Meal planning that actually knows your kitchen.
Most meal planners ask you what you want for dinner and search a generic database. DinDin writes meals around four things only your household can tell us.
- Your household
Picky eaters, allergies, ages, spice tolerance. The system writes each recipe with the specific people at your table in mind.
- Your kitchen
Appliances you own, pantry items on hand, ingredients you keep stocked. Recipes get steered toward what you can actually cook tonight.
- Your calendar
What you cooked last week, what's set to repeat, what to skip. The plan never repeats yesterday's dinner or recommends something you just had.
- Your taste
Favorites, cuisines you lean into, seasonality. Suggestions get more accurate every time you favorite or pass on a recipe.
Pillar 1
Tailored AI generation
When you ask for new recipes, the generator receives your full profile as context — not just a cuisine and a calorie target. Household members are named in the prompt. Allergies are hard constraints. Available appliances steer technique selection.
- Cooking for Eli, age 4 — picky, no spice. Mia, age 7 — adventurous.
- Used your air fryer because it's faster on weeknights.
- Avoiding cilantro per Jeffrey's profile.
- No repeat from last 14 days (last salmon dish: 5 days ago).
- Targeting 35-minute total time. You favored two Thai recipes recently.
You see this happening live — the streaming "Considering…" widget on the recipe generator surfaces every context item the AI consulted while it writes.
Pillar 2
Smart course suggestions
When you schedule a main, the system already knows enough to offer the right sides, appetizers, and desserts. Same context the dinner generator used flows into the course generator — no separate setup, no re-entering preferences.
- Cuisine is American. Total prep + cook is 22 minutes.
- You have jasmine rice on hand. Pantry shows soy + sesame oil.
- No kid-friendly side scheduled this week.
- Last linked side for salmon was a slaw — try something starchier this time.
- Dessert: lean fruit-forward, the meal is already rich.
The first time you save a linked course on a dinner, the system also offers to remember the link so next time you plan that dinner the side appears automatically.
Pillar 3
Drink pairings worth pouring
Pairing suggestions for wine, beer, cocktails, and zero-proof alternatives — each with the flavor profile, alternatives, and (for cocktails) a recipe. The system respects your alcohol preferences and never suggests something you'd skip.
- WINE — Albariño (bright, mineral, citrus). Also try: Vermentino, Pinot Grigio.
- BEER — Witbier (light, herbal). Also try: Belgian blonde, Pilsner.
- COCKTAIL — Margarita. Also try: Paloma, Caipirinha.
- MOCKTAIL — Sparkling lime soda with cilantro syrup.
Cocktails come with ingredient lists and method so you don't have to look anywhere else.
It gets better the more you use it.
DinDin learns from every favorite, every cooked dinner, every skipped suggestion. Your second week is better than your first; your second month is better than your first.
- Recipes you cook get prioritized. Recipes you skip surface less often.
- Patterns are detected. If you've made pasta four Mondays in a row, the Pattern Breaker hero surfaces a new direction.
- Recipes you mark "don't repeat" stay out of the automatic recycle queue forever.
- Drink pairings adapt to what you've actually poured — the system stops suggesting beer if you only ever pick wine.
Your data stays yours.
Export your entire library at any time. Cancel any time and we delete what you ask us to delete. We never train models on your private data. The personalization above only happens inside your account — it never feeds anyone else’s plan.