There are more "AI meal planner" apps than ever right now. Ads for them are everywhere. They all promise to solve the dinner question forever.
Most of them are not actually using AI in any meaningful way. They're recommendation engines with an AI badge on the app store listing. And the ones that do use real AI often limit it in ways that make it much less useful for real families with real constraints.
We tested every significant option. Here's the honest breakdown — and why the answer might surprise you.
Why most "AI meal planner" apps aren't really AI
The term "AI" on an app listing usually means one of three things:
- A filtering algorithm — the app matches your dietary preferences against a fixed recipe database. It's not generating anything; it's sorting.
- A basic recommendation engine — it tracks what you've made before and suggests similar recipes. Useful, but not the same as AI reasoning.
- Actual generative AI — the app uses ChatGPT, Claude, or a similar language model under the hood and generates personalised meal suggestions. This is rare and usually only available at higher price tiers.
The test question: Can the app handle this instruction? "Plan 5 dinners for a family of 4. My 7-year-old won't eat mushrooms or visible onion. Budget is $100. Use whole food ingredients — no artificial additives. Each meal under 40 minutes." If the app can't take that instruction and produce a useful, specific plan — it's not using real AI.
We put every app through exactly this test. The results were stark.
The criteria that actually matter for families
Before the comparison table, here are the criteria we used. These are the things that actually break or make a meal planning tool for a family with kids.
- Household understanding — can it remember your family size, budget, and picky eater rules without re-entering them every week?
- Real food support — does it plan whole food, additive-free meals by default, or does it push processed shortcuts?
- Budget flexibility — can you give it a hard dollar limit and have it plan within that number?
- Picky eater handling — can it accommodate specific exclusions (not just broad dietary categories) for individual family members?
Full comparison: the tools we tested
Six tools, tested over four weeks with the same household: family of 4, one picky 7-year-old, $100/week grocery budget, preference for whole food meals under 45 minutes.
| Tool | Pricing | Real food? | Budget control? | Picky eaters? | Household memory? |
|---|---|---|---|---|---|
| ChatGPT (with prompts) | Free / $28/mo | Yes (with rules) | Yes | Yes | Paid only |
| Claude (with prompts) | Free / $30/mo | Yes (with rules) | Yes | Yes | Projects (paid) |
| Ollie | ~$15–20/mo AU | Partial | No | Basic only | Profile only |
| Mealime | Free / $6/mo | Partial | No | Category-level | Profile only |
| Prepear | Free / $8/mo | Partial | No | No | No |
| Reciptz | Free / $10/mo | Partial | No | No | No |
ChatGPT (free and paid)
The most flexible option by a wide margin. With the right prompts, ChatGPT handles every constraint a real family throws at it — budget limits, whole food rules, picky eater exclusions, time constraints, all at once. The free plan works for weekly planning sessions. The paid plan ($28/month AU) adds memory so you don't re-paste your household rules each week.
Verdict: Best overall for customisation and whole food quality. Slight learning curve with prompts. See our step-by-step ChatGPT meal planning guide for the exact prompts.
Claude (free and paid)
Marginally better than ChatGPT at following precise, detailed instructions — which matters when your household rules are specific. The paid plan's Projects feature stores your rules permanently, so setup is genuinely one-time. Claude is the AI underneath our Meal Planning OS for exactly this reason.
Verdict: Best if you want permanent household memory. Slightly less familiar to most users than ChatGPT but outperforms it on instruction-following.
Ollie
The most polished dedicated app. Clean interface, recipe cards with nice photography, good grocery list functionality. But it's built for individuals and couples, not families with kids. Picky eater handling is limited to broad dietary categories (no dairy, no gluten) rather than specific ingredient exclusions. No budget control. And at $15–20/month, it's hard to justify when the free AI tools do more.
Verdict: Great UI, limited for family complexity. Better suited to a household without kids or strong food preferences.
Mealime
A solid recipe app with a meal planning feature. Good recipe quality and a clean grocery list. But the "AI" element is a filter, not generative — it can't handle specific constraints like "my 7-year-old refuses mushrooms and visible onion." Budget control is absent. The free plan is usable; the paid plan adds more recipes but not more intelligence.
Verdict: Good for recipe discovery, not for household-specific planning. Best used as a recipe library alongside a real AI planning tool.
Prepear and Reciptz
Both are recipe organisation tools with meal planning grids. Neither uses generative AI. You're essentially pinning recipes into a calendar. Useful for organisation but not for solving the "what do I cook this week?" decision from scratch.
Verdict: Organisational tools, not AI planners. Don't pay for these when free options exist.
The clear winner — and why
Using ChatGPT or Claude directly — with your own household rules built into the prompt — outperforms every dedicated app for most families. Here's why.
Dedicated apps have fixed recipe databases and limited customisation. They can filter, but they can't generate. When your family has a specific set of constraints — budget, ingredient standards, picky eater rules, time limits — a fixed database runs out of good matches quickly. You end up seeing the same recipes on rotation.
ChatGPT and Claude generate a fresh plan every time, working within whatever constraints you give them. That means the plan is always new, always specific to your household, and always within your rules. Not a filtered subset of someone else's recipe library.
Ready to run your first ChatGPT or Claude meal plan? Start with the free prompt pack.
Grab the free prompts →The one thing dedicated apps can't do
Real household memory and customisation at the level a family actually needs.
Every app we tested uses a profile system — you set dietary preferences (vegetarian, gluten-free, etc.) and the app filters accordingly. That's fine for broad categories. But a real family's constraints are much more specific.
"My 7-year-old won't eat mushrooms, visible onion, anything with a strong sauce, or anything that looks green" is not a dietary category. It's a household rule. Apps can't store that and reason from it. ChatGPT and Claude can.
"We spend $100 a week on groceries, we prefer to share ingredients across meals to reduce waste, and we always try to batch cook something on Sunday that becomes at least two weeknight dinners" — that's a planning strategy, not a filter. No app handles it. AI does.
This is the core gap the Meal Planning OS closes: it stores all of your household rules — picky eater specifics, budget, ingredient standards, your family's favourite meals, cooking style — in a Claude Project. Then "plan" is all you have to type. The AI knows everything else.
When a dedicated app IS the right choice
Apps aren't useless. There are specific situations where they genuinely win.
- You want recipe cards with photos — AI generates text, not formatted recipe cards. If you want something that looks like a cookbook page while you cook, Mealime or Ollie beats a ChatGPT text block.
- You're not comfortable with prompts yet — apps have a simpler onboarding. If building a prompt feels like too much friction right now, start with an app and migrate later.
- Your household has standard dietary categories — if your only constraint is "vegetarian" or "gluten-free" with no other complexities, a filtered recipe app works fine.
- You want to save and organise recipes — for building a personal recipe library from things you've cooked and loved, apps like Prepear are actually useful.
For most families with real complexity — kids with preferences, budgets to hit, whole food standards to maintain — the AI approach wins. But if you're just starting out and want something with less setup, Mealime's free plan is a reasonable place to begin.
The hybrid approach: Some mums use both — ChatGPT for weekly planning (what to cook), and Mealime or a recipe app for the how-to while cooking. The AI handles decisions; the app handles recipe cards. That combination costs nothing and works brilliantly.