Browse the blog index, cook mode + YouTube, Georgian supra piece, and download Suphra.
There are four people in the household. One wants pasta. One is "not really feeling carbs right now." One doesn't care as long as it's not soup. One is nine years old and has opinions about textures.
This is Tuesday. This is most households. And this is the problem that Suphra's 3+3 recipe roulette was built for: not one suggestion, but six genuinely different directions — before you change a single ingredient.
Why one recipe suggestion is never enough
Every major recipe platform defaults to a single recommendation. You search or input, you get one answer. If that answer doesn't work — wrong cuisine, wrong effort level, wrong vibe — you're back to searching. According to Tastewise's 2026 food trends analysis, the #1 driver of food decision abandonment is friction in the selection process — not lack of options, but lack of quickly comparable options.
Suphra generates three recipes side by side, each showing difficulty, time, and core ingredients at a glance. If none of them land, you hit regenerate — three more, same ingredients. Six directions total before you've touched a single input.
The "weird pantry" problem: scan-first keeps the roulette honest
Recipe suggestion tools that don't start from your actual inventory have a fundamental flaw: they suggest things you can't make. You get excited about a recipe, tap through to the ingredients, and realize you'd need to go shopping. The suggestion is useless.
Suphra's scan-first approach means the roulette only spins on what's real. Photograph your fridge. Scan your pantry. Add whatever's on the counter. Every recipe generated — all six of them — is cookable from exactly what you scanned. No phantom ingredients. No emergency shopping runs.
Research from Cornell University's Food and Brand Lab found that presenting people with multiple pre-filtered options — rather than open-ended search — significantly reduces decision fatigue and increases follow-through on meal preparation. Curated choice, not infinite scroll, is what gets dinner on the table.
Cornell Food and Brand LabWhen nobody agrees: the picky household use case
Picky eating isn't just a child behavior — IFT's 2026 food trend data shows that "flexible eaters" — people who shift between dietary preferences depending on mood, energy, or social context — now represent the majority of adults. Monday's plant-based person is Friday's steak-craving person. Tuesday's pasta night becomes Thursday's "I'm not really feeling carbs."
Six distinct recipe directions from a single pantry scan means something for nearly everyone. The 3+3 format is built specifically around this: three attempts, then three more, before you've added or removed a single ingredient. Enough range to find the overlap between a nine-year-old's texture opinions and a partner's protein goals.
Regenerate without guilt: why more options is the right design
There's a subtle anxiety in single-recommendation systems: if you reject the suggestion, you feel like you're doing something wrong. You second-guess the rejection. You make the suggested meal anyway, even though nobody wanted it.
Suphra's regenerate button removes this. The system expects you to use it. Three options is a starting point, not a verdict. Hit regenerate, get three more, pick the one that fits tonight. It's the same freedom as scrolling a menu at a restaurant — without the $18 salads and the 45-minute wait.
Scan your protein first — it's usually the limiting ingredient. Suphra will build all six recipes around making the most of it, so whatever you have — one chicken breast, a can of chickpeas, leftover mince — becomes the anchor rather than the afterthought.
Weird pantry? Picky household? Hit spin.
Scan what you have, get three recipes side by side, regenerate for three more. Six cookable options in under 30 seconds — from whatever's actually in your kitchen.