To use the What Should I Eat Generator, set any combination of meal, mood, and time filters, draw a suggestion from the filtered pool, then verify the idea against your own ingredients, dietary needs, and recipe choice. The widget works from a transparent editorial pool of exactly 24 original food ideas, organized as four meal values crossed with three mood values with two ideas in every combination. Every draw happens locally in your browser using Web Crypto randomness with rejection sampling, and the current result is excluded from the next draw when another eligible idea exists, so two consecutive clicks never return the same entry. Nothing is uploaded or saved, and the tool never asks for your location, pantry, allergies, or previous meals. What you get back is a title, a short direction, a broad time label, and the count of ideas that survived your filters — not a complete recipe, an ingredient list, or a nutritional verdict. That handoff is deliberate: the What Should I Eat Generator solves ordinary meal indecision, while you make the real cooking decision.
Inside the Generator's 24-Idea Pool
The generator draws from a fixed list of 24 ideas written specifically for the tool, not pulled from a restaurant or recipe database or any cuisine authority. The structure is a clean grid: four meal values (breakfast, lunch, dinner, snack) crossed with three mood values (fresh, cozy, hearty), with exactly two ideas reserved for each of the 12 cells, so every meal-and-mood combination has a guaranteed pair of suggestions. Breakfast ranges from fruit and yogurt to a potato hash. Lunch includes salads, soups, noodles, bowls, and sandwiches. Dinner covers tacos, pasta, rice, stir-fry, and a baked potato. Snack ideas include fruit, crackers, popcorn, toast bites, a folded quesadilla, and pantry snack mix.
You can leave any filter blank to widen the pool, but the underlying inventory never grows or shrinks — the filters only narrow what you see drawn. The widget's job is narrow on purpose. It returns a single idea from the pool that survives your filters, then lets you copy it for a shopping note or use it as a prompt for a recipe search. It does not rank local businesses, calculate nutrition, compare prices, or look at your previous meals. Knowing that boundary up front makes the rest of the workflow easy to trust, because the tool only promises what its contract actually delivers.
How to Use the What Should I Eat Generator
The full workflow is short and repeatable. Each step produces a piece of information you need before deciding whether the suggestion fits your situation, and the same five steps cover every draw you will ever run.
- Open the What Should I Eat Generator and pick a meal — breakfast, lunch, dinner, or snack — or leave the meal filter open if you are open to any of them.
- Pick a mood — fresh, cozy, or hearty — and optionally set a rough maximum preparation time of 15, 30, or 45 minutes. Leave any filter open if a larger idea pool suits you better.
- Read the eligible-pool count so you know how many ideas are still in play after filtering, then draw a suggestion from the remaining entries.
- Review the title, the short editorial direction, the rough time estimate, and the eligible-pool count returned with the result.
- Verify the ingredients, allergens, dietary needs, food safety, and recipe before you cook or copy the idea. If anything is off, draw again or widen a filter and redraw.
Reading the Meal, Mood, and Time Filters
The three filters do different work. The meal filter caps the kind of occasion the idea is tuned for, the mood filter nudges the texture and warmth of the result, and the time filter limits how much active effort the rough planning label allows. Leaving any of them open simply means "no constraint on this dimension," which enlarges the eligible pool before the draw. The labels themselves are practical interface categories rather than rules — a breakfast wrap can be dinner, and a snack mix can become part of lunch when it fits the situation.
| Filter | Options | What it controls |
|---|---|---|
| Meal | Breakfast, lunch, dinner, snack, or open | Which of the four meal groups the eligible pool is drawn from. |
| Mood | Fresh, cozy, hearty, or open | Texture and feel — crisp produce and brighter combinations, warm familiar food, or more substantial formats. |
| Time limit | 15 min, 30 min, 45 min, or open | Maximum rough preparation estimate allowed in the eligible pool. |
Time labels are rough planning estimates chosen by the tool, not guarantees of completion time. They exclude shopping, thawing, marinating, unfamiliar techniques, cleanup, and interruptions, and they assume common ingredients and basic equipment are already on hand. Actual preparation can be shorter or longer depending on what you bring to the kitchen. A time label is also not a food-safety cooking time, so a 15-minute suggestion is not a claim that the food is safe to eat in 15 minutes.
How the Random Draw Is Produced
The draw is built to be auditable. Once filters have narrowed the pool, the browser asks the Web Crypto API for an unsigned 32-bit word via crypto.getRandomValues. To prevent the slight bias that a naive modulo would introduce when 2^32 is not an exact multiple of the pool size, the tool rejects any word that lands at or above the largest multiple of the pool size that fits inside 2^32, then retries until an accepted value arrives. The accepted word is then mapped to a position by modulo, giving every eligible entry the same probability.
For a no-repeat draw with a remaining pool of 7 (after one idea was excluded from a pool of 8), the boundary math works out as follows:
- 2^32 = 4,294,967,296.
- floor(4,294,967,296 / 7) = 613,566,756.
- 613,566,756 × 7 = 4,294,967,292.
- The 4 values from 4,294,967,292 up to 4,294,967,295 are rejected; any value below 4,294,967,292 is accepted.
- The accepted word is reduced modulo 7 to land uniformly on positions 0 through 6.
Retries are bounded — at most 128 attempts — so a misconfigured or starving source cannot loop forever. When the pool has more than one entry and a previous result exists, that previous ID is removed from the next draw only, which is why the full eligible count stays visible to you while the very next click is guaranteed to differ. Later draws can return the same idea again once the no-repeat window has closed.
What Each Result Includes — and What It Doesn't
Every draw returns four pieces of information: a title, a short editorial direction, a rough preparation-time label, and the eligible-pool count that survived your filters. That is the complete output. The widget does not attach a calorie total, a protein figure, an ingredient list, a cooking method, or a substitution guide, because those would imply a level of certainty the editorial copy cannot stand behind.
| Returned by the tool | Left to you |
|---|---|
| Title and short direction | Choosing a complete, trusted recipe |
| Rough preparation-time label | Verifying allergen and dietary safety |
| Eligible-pool count for the active filters | Confirming ingredients are on hand |
| Copy-idea action for notes or recipe prompts | Adjusting portions, cooking temperature, and storage |
This split keeps the random step honest. The generator can promise a fair, local, repeat-resistant draw from a transparent pool, and you keep the decisions that depend on your pantry, your culture, your schedule, and the people eating the meal. Clipboard access is requested only when you press Copy idea, so the rest of the workflow leaves no trace on your device.
When to Draw Again or Widen a Filter
If a result does not match your situation, the cleanest response is another draw from the same filters, because the current ID has already been excluded from the next pick. If the eligible count drops to one, widening a filter is the right move — drop the mood, lift the time limit, or open the meal — until at least a handful of ideas are back in play. There is no penalty for widening. Filters only narrow the eligible pool, so relaxing them does not weaken the random step; it just gives the next draw more positions to choose from.
If the entire pool becomes empty because of an over-tight combination, the tool returns no result without requesting any entropy, so you will not see a placeholder or a stale pick. Widen at least one filter, watch the eligible count rise, and draw again. The full inventory is small and transparent, so when you widen a filter you can predict roughly how many ideas will re-enter the pool — but the actual draw is still chosen by the same Web Crypto step described above.
Verifying Ingredients, Allergies, and Food Safety
The tool intentionally has no dietary checkbox, because a dish name such as tomato soup, yogurt bowl, stir-fry, or quesadilla does not establish which ingredients, allergens, preparation surfaces, or substitutions will be used. The same named dish can be suitable or unsuitable depending on the recipe, so the generator does not label ideas vegan, vegetarian, gluten-free, halal, kosher, low-sodium, diabetic-friendly, pregnancy-safe, or allergy-safe. Before you cook, confirm that the meal and rough effort fit, choose a recipe you trust, and check every ingredient and preparation method against the needs of the people eating it.
For medically prescribed diets, serious allergies, infant feeding, pregnancy concerns, swallowing problems, or any other health-related need, rely on qualified guidance and verified ingredient information rather than a random prompt. The generator's best use case is ordinary low-stakes indecision: you want a broad idea, you understand your own constraints, and you will make the real food decision yourself. Treat the suggestion as a conversation starter, and let a trusted recipe and your own judgment finish the job.