ChatGPT Prompt for Weekly Meal Planning: Free Template (with Aldi/Lidl/Rewe Deals)
ChatGPT can build a weekly meal plan from supermarket deals only if you paste the current offers into the prompt yourself. Without real prices, the model invents numbers. The copy-paste template below gives the model the right structure: a list of deals, your household constraints, a daily plan, shopping list, and estimated cost.
If you don't want to scan deals every week, a specialized app does that for you. More below.
Why the obvious question doesn't work
When people type into Google or directly into ChatGPT — "Can you help me create a weekly meal plan based on supermarket offers?" — they get a polite but useless answer. The model suggests generic dishes, invents arbitrary prices, and knows nothing about the current Aldi Süd promotions or this week's Lidl Hello range.
That isn't a model intelligence problem. ChatGPT simply has no live access to supermarket flyers. Its training data ends months before the cutoff date, and a "chicken breast €0.99/100g" from the training set might be €1.39 today.
The fix is to feed the model the data it doesn't have — as part of the prompt itself.
The copy-paste template
This prompt works equally well with ChatGPT (GPT-4 or newer), Claude, and Gemini:
You are my weekly meal planner. Build a 7-day plan for [PEOPLE] people based on
the supermarket deals listed below.
Constraints:
- Budget: [BUDGET] € for the whole week
- Diet: [e.g. omnivore / vegetarian / vegan / gluten-free]
- Allergies/no-gos: [e.g. no nuts, no mushrooms]
- Cooking time per evening: max [MINUTES] minutes
- Leftover day: [yes / no]
Current deals this week:
- Aldi: [item] [quantity] [price], [item] [quantity] [price], …
- Lidl: [item] [quantity] [price], …
- Rewe: [item] [quantity] [price], …
Please return:
1. A Monday–Sunday daily plan with breakfast, lunch, and dinner
2. A consolidated shopping list, grouped by store
3. Estimated total cost per day and for the week
4. Three optional leftover-use ideas
Ingredients may repeat across days. Plan deliberate overlap so nothing spoils.
No ingredient should appear in the plan that isn't in the deal list or in the
pantry staples (salt, pepper, oil, flour, rice, pasta, eggs, onions, garlic).
Fill in the placeholders. Setup time: about five minutes — assuming you've already gathered the deals.
What actually matters in the prompt
1. Concrete offers with prices. Without prices, the model hallucinates. Even when it suggests good dishes, the cost calculations are random. Rule: if a price isn't in the prompt, it's made up.
2. Clear constraints. People and budget are mandatory. Cooking time, diet, and no-gos are optional but lift output quality noticeably.
3. Structured output request. The "1. plan, 2. list, 3. cost, 4. leftovers" enumeration forces the model into a predictable shape. Without it, you often get prose answers you can't act on directly.
4. Anchor for non-listed ingredients. The pantry-staples line at the end stops the model from putting exotic ingredients into the plan that you don't have at home.
Sample input and output
A filled-in prompt for a two-person household looks like this:
People: 2
Budget: 50 € for the week
Diet: omnivore, lots of vegetables
Allergies: none
Cooking time: max 30 minutes
Leftover day: yes
Aldi deals this week:
- Chicken breast 1 kg €6.99
- Bell peppers 500 g €1.49
- Plain yogurt 500 g €0.89
- Potatoes 2.5 kg €1.99
- Basmati rice 1 kg €1.79
Lidl deals this week:
- Eggs 10 ct €1.99
- Frozen spinach 750 g €1.49
- Tomatoes 500 g €1.29
Output typically includes a chicken-rice skillet, a yogurt bowl, a spinach-potato bake, a tomato salad, an egg pancake, and a leftovers night — with portions per person, a sorted shopping list, and a plausible total of €38–€48.
The underlying method isn't new. It follows the system from our Build a weekly meal plan from supermarket deals post — only this time the language model handles the "plan composition" step.
The three most common mistakes
1. Pasting deals without quantities
"Chicken breast €6.99" is ambiguous. 200 g or 1 kg? Without a quantity, the model picks a plausible-looking number — and you end up with too much or too little protein in the cart. Always include quantities.
2. Not double-checking the plan
Language models are confident. Even when they double-count an ingredient or consolidate the shopping list incorrectly, the answer sounds authoritative. Reconcile list against plan once before shopping.
3. Re-writing the prompt every week
ChatGPT forgets everything between weeks. Saving the prompt as a Custom GPT or a text snippet saves five minutes of typing weekly. Tip: tune the pantry staples list once to your household and never touch it again.
Where ChatGPT still hits a wall
Even with the perfect prompt, the core problem stays: you have to research the deals yourself. Realistically, every Sunday evening:
- Walk through the Aldi Süd flyer
- Walk through the Lidl Hello flyer
- Walk through the Rewe leaflet
- Type or paste 4–8 relevant items per store
That's the actual work. The prompt only handles the composition step.
For more on the manual workflow without AI, see our posts on Recipes based on sales and Grocery shopping by recipe.
Where a specialized app beats ChatGPT
Flyva was built specifically for the step ChatGPT can't do: automatic ingestion of current discounter flyers. The app knows the live Aldi, Lidl, Rewe, and Edeka offers in your region and combines them with matching recipes into a finished weekly plan and shopping list. Flyva is available free on the App Store and Google Play — no invite, no waitlist.
You enter household size, budget, and diet once — and get a complete plan every Monday. No prompt, no deal research, no fact-checking. More background in our post on meal planning apps and grocery savings and in the step-by-step guide for next week's meal plan, which shows what the prompt approach versus the app approach actually save.
When each tool makes sense
| If you … | then … |
|---|---|
| enjoy optimization and rewriting prompts every 1–2 weeks | ChatGPT with a solid template is ideal |
| already scan the flyers and just want help composing | the template above is perfect |
| want a plan every week without effort | a specialized app like Flyva fits better |
| have complex diets (allergies, keto, halal) to enforce | an app with diet filters is more reliable |
Skip the prompt and start cooking
If prompt engineering isn't your idea of a Sunday evening, these guides give you ready-to-use plans and recipes:
- Aldi weekly meal plan — full sample 7-day week
- Lidl weekly meal plan — Lidl-specific variant
- Rewe weekly meal plan — Rewe variant
- Edeka weekly meal plan — Top deals and Marktfrisch
- Penny weekly meal plan — Bio Helden + PAYBACK
- Kaufland weekly meal plan — bulk packs for families
- Budget vegetarian weekly meal plan — meatless under €35/week
- Single-person weekly meal plan — for one under €25/week
- Quick budget recipes — German classics — fallback options
- Student meal plan on a budget — small budget, big result
- Lentil bolognese recipe — the unbeatable anchor
- Smart ways to save money on groceries — companion read
The full collection of deal-friendly meals lives in our recipe database — every recipe links to the cheapest available store.
Bottom line
A good ChatGPT prompt for a weekly meal plan rests on three things: concrete deals, clear constraints, structured output. The five-minute setup pays off immediately.
If you want to skip the deal-research step too, a specialized app gets you there faster. Both approaches work — the only question is how much of the work you want to do yourself.
Frequently asked questions
Can ChatGPT create a meal plan based on supermarket offers?
Only with help. ChatGPT does not know current Aldi, Lidl, or Rewe prices. You have to research the deals yourself and paste them into the prompt. With that input, ChatGPT can produce a usable weekly plan. The template in this article shows the exact format.
What prompt works best for a weekly meal plan?
Three parts matter: first, the actual deals as a list with prices; second, your constraints (people, budget, diet, allergies); third, an explicit request for structured output (daily plan + shopping list + estimated cost). The full template is in the article.
Why does ChatGPT sometimes invent prices?
Because the model never saw discounter flyers during training and hallucinates prices when none are in the prompt. Rule of thumb: if a price is not in your prompt, it is made up. For realistic plans, always include the actual deal prices.
Is a specialized app better than ChatGPT?
If you enjoy doing the research weekly, ChatGPT with a good prompt is fine. If you want to skip the deal-scanning step, a dedicated app like Flyva is the better fit — it ingests current flyers automatically, knows the prices, and delivers the plan in under five minutes.
Does the prompt also work with Claude or Gemini?
Yes. The template is model-agnostic. Claude 4.x and Gemini 2.5 produce comparable or slightly better results because they generate structured output a bit more reliably. The prompt logic is the same.