CookSnap vs. SuperCook: Why We Built a Different Kind of Recipe Finder

If you have ever searched “app to find recipes from ingredients I have,” you have found SuperCook. They were the first to do this well, and for a long time they were the only ones. Anyone building in this space — including us — owes them a debt for proving the category exists.

We get asked, several times a week now, what we do differently. So here is the long answer, with the parts where SuperCook wins kept in.

What SuperCook does well

SuperCook’s superpower is scale. They index millions of recipes scraped from across the web. If your pantry has thirty items in it, SuperCook will find a match. They built an ingredient autocomplete that handles the long tail beautifully — “haricots verts,” “pomegranate molasses,” whatever weird thing is on your shelf, it’s in there. They also handle “pantry staples” (salt, pepper, oil) as a toggle so you don’t have to retype them every time. That is a genuinely thoughtful UX detail and we copied the pattern in spirit if not in exact form.

Their recipe corpus is broader than ours and probably always will be. If you cook serious volume across many cuisines and want every possible match, SuperCook is the better fit.

Where the approaches diverge

The friction with SuperCook, in our experience and in the reviews we read every week, is downstream of where the recipes come from. Their library is a wide-net scrape from food blogs, community sites, and content farms. That gives them volume. It also means a single “chicken and rice” query returns results where:

• The instructions are written for a different climate or altitude.
• The author skipped key steps because the blog post had a backstory.
• The ingredient list has been rewritten by an SEO tool and quantities are vague.
• You click through and discover the recipe is gated behind ads, a video, or a newsletter signup.

SuperCook is not at fault for any of this — they index the open web and the open web is what it is. But the user experience suffers. You typed five ingredients and got a recipe that asks for three more and a thirty-minute prep you can’t verify.

CookSnap takes the opposite trade-off. Our library is ~9,000 recipes, curated and vision-verified, every single one. Hero images are either generated or sourced through a strict pipeline. Ingredients are normalized to a canonical taxonomy. Steps are checked for completeness. The library is smaller. The match quality is higher.

The recipe-source axis

We think of recipe apps as living on a spectrum, with two endpoints:

1. Maximum breadth. Scrape the open web. Index everything. SuperCook lives here.
2. Maximum verification. Curate, normalize, hand- check. CookSnap lives here.

Pure-LLM apps (DishGen, FoodsGPT, ChefGPT) sit off the chart entirely — they generate recipes that never existed before your query. That is a different problem with a different set of failure modes; we covered it in a separate post.

Three concrete differences a user will notice

1. Photo-based ingredient input.CookSnap’s iOS app uses on-device vision to identify ingredients from a photo of your fridge. SuperCook requires manual selection from a list. Typing thirty things is the part most people give up on; the camera removes that friction.
2. Fit percentage that means something.Every CookSnap match shows you a fit percentage based on canonical ingredient overlap and missing-staple count. SuperCook tells you which ingredients are missing but doesn’t score the match, so a 90% match and a 40% match look the same in the results list.
3. One destination, not a click-out.Tap a CookSnap result and you get the full recipe on our domain — no ads, no autoplay video, no newsletter wall. SuperCook is fundamentally a search engine for other people’s recipe sites, so every match is a click-out into a third-party blog.

When SuperCook is the better choice

We mean this honestly: if you have a deep pantry and you like bouncing between cuisines, SuperCook will surface more options. Their long-tail ingredient coverage is excellent. If your priority is “most possible matches,” pick them.

Where we win is the seven-thirty-pm-with-a-tired-toddlerscenario. Four ingredients. One recipe. Cook now, no clicking through to a food blog, no figuring out which of the seventy results is the one that won’t fail. That’s the problem we are obsessively focused on, and that focus is what shapes every product decision we make.

Try them both

If you’ve never used either, run the same five ingredients through SuperCook and CookSnap’s free recipe finder. Compare the top three results from each. Notice which ones you could actually start cooking right now, and which ones you’d bookmark and forget. That diff is the design philosophy.