Help & FAQ

Nutty is your personal food-decision assistant. You describe your situation, whether that's a diagnosed condition, symptoms without a diagnosis yet, or just wanting to eat less inflammatory food, and Nutty derives a starting protocol for you. From there, you photograph ingredient labels and Nutty tells you if a product fits your protocol, with plain-language reasoning and a swap suggestion when it doesn't.

No. Nutty never diagnoses, treats, or guarantees an outcome. Everything it gives you, your protocol, a scan verdict, a recipe, is a wiser-choice suggestion grounded in an anti-inflammatory framework, not a clinical judgment. If you're managing a diagnosed condition, use Nutty alongside your doctor or dietitian, not instead of them.

Onboarding asks about your actual situation (a named diagnosis, symptoms without one, or just eating better) and any specific conditions or medications that affect what to flag. From that, Nutty derives a starting protocol, either the everyday anti-inflammatory baseline or a fuller strict AIP elimination, whichever fits what you told it.

Yes. Open Protocol from the bottom nav and use "Retake Onboarding" to update your situation, conditions, or medical flags. Your day-to-day exceptions (things you react to or tolerate), and the "Anything else Nutty should know" free-text box, can both be edited directly on the Protocol page any time, without retaking onboarding.

Use the "Anything else Nutty should know" box on your Protocol page. Write it in your own words, no diagnosis required, whether you're watching your sugar, keeping an eye on your blood pressure, or anything else about how you eat. Nutty reads it every time it checks a label, and steers its swap suggestions and cautions accordingly. It never changes whether a product fits your core anti-inflammatory protocol, that part stays consistent no matter what you write here.

You photograph an ingredient label, and that photo, along with your current protocol, goes to Nutty. Nutty reads the ingredients and returns a verdict (fits your protocol, doesn't fit, or couldn't read it clearly), the reasoning behind it in plain language, and a swap suggestion when something doesn't fit.

Nutty is careful, but it's reading a photo, not a lab. Blurry labels, glare, or unusual fonts can affect what it catches, that's what the "couldn't read it clearly" verdict is for. Use your own judgment for anything you're unsure about, and treat every verdict as a starting point, not a guarantee.

Yes. Open any scan's details and use "Ask Nutty" to ask something specific about that exact label, right when you scan it or later. Every scan saves its own photo automatically, so a follow-up works whether you ask it right away or come back to it another day.

Yes to both. Every scan is saved automatically to your Scan History, verdict and photo together, so you can revisit it or ask follow-up questions later. You can rename any scan or delete it entirely, any time, from Scan History, deleting a scan removes its saved photo too. Nobody but you can see your scans.

Packaged food has to list its ingredients. Restaurants don't. A menu names a few things, but not the oil, the marinade, the flour, or the sauce. So Nutty can't read a menu the way it reads a label. It tells you what the menu says, what's likely, and what to ask your server. The questions are the important part.

Likely fits: nothing on the menu conflicts with your protocol, though you can still ask if you're unsure. Ask before you order: it could work, but something the menu doesn't say could change that, so ask the questions first. Probably not for you: the menu names something your protocol skips. None of these is a guarantee. Treat them as a starting point.

A "GF" mark is the restaurant's own claim. Flour dusting, thickened sauces, and shared cooking surfaces can still add gluten, and menus rarely say so. So for fish and meat, Nutty always suggests asking, and for GF-marked dishes it words the question as confirming the claim. Asking costs you a few seconds.

No. The photo is read for that one request and not kept. What is saved is Nutty's notes on each dish and anything you write in Menu History. You can delete any menu, and your notes with it, at any time.

Every menu you scan is saved to Menu History. After you eat, you can write down what the server told you, what you asked them to change, and whether the dish worked for you, or that you didn't try it. The "Worked for me" tab collects every dish that did, so you know what to order again and how to ask for it. You can rename or delete any menu, any time. Nobody but you can see them. This is food guidance for your protocol, not medical advice.

You can build a recipe from ingredients you already have, from a craving, or adapt one to better fit your protocol, Nutty generates it and you save the ones you want to keep under My Recipes.

A separate library of recipes written directly by the founders, Chef Gerald DeBerry and Jeanine Grace. It's browsable the same way as My Recipes, but it isn't something you edit yourself.

Every saved recipe belongs to a category you choose. Use "Move" on a recipe to shift it to a different category (or type a brand-new one), and "Rename" on a category heading to rename every recipe in it at once. Delete a recipe any time from the same screen.

One plan, $9.99/month, everything in the app included, no separate tiers and no free version. Billing is handled entirely by Stripe.

Open Account and tap "Manage Subscription". That takes you to Stripe's own billing portal, where you can update your payment method, view invoices, or cancel. Cancelling stops future charges at the end of your current billing period; the app itself doesn't hold a separate cancellation flow, Stripe's portal is the real one.

Charges for the current billing period aren't refunded once made, but cancelling any time stops all future charges. If something went wrong with a specific charge, email us at support@healthnutai.com and we'll look at it directly.

Open Account from the bottom nav and tap "Delete my account." You'll see what gets removed, then type DELETE to confirm. It happens right away: your protocol, scans, menus, and saved recipes are all deleted, and your subscription is canceled that day, with no more charges. Days already paid for aren't refunded. Stripe keeps its own record of past payments. You can also email support@healthnutai.com from the address on your account and we'll do it for you.

Yes. Your protocol, scans, and recipes are yours alone, enforced at the database level, not just hidden in the app's screens. No other user, including us, can browse your data through the app.

Yes, by necessity, that's how Nutty reads your labels and reasons about your protocol. Your scan photos and protocol details are sent to Anthropic (the company behind Claude, the AI Nutty runs on) to generate a result. See our Privacy Policy for the full picture.

Both are linked at the bottom of this page, and from the sign-up screen.

No. Health-Nut AI is a Progressive Web App (PWA), it runs straight from your browser and can be added to your phone's home screen so it opens and feels like a regular app, with no App Store or Google Play install required.

On iPhone (Safari): tap the Share icon, then "Add to Home Screen." On Android (Chrome): tap the ⋮ menu, then "Add to Home screen" or "Install app." Either way, it'll open in its own window from then on, just like an installed app.

Check that Camera access is turned on for your browser: go to iPhone Settings, find your browser (Chrome, etc.) in the app list, and make sure Camera is enabled. That fixes it most of the time. If it's already on and still black, try opening Health-Nut AI in Safari instead, or use "Upload a Photo" to pick an existing photo or take one through your phone's own camera app.

Yes, though it's your phone's keyboard doing it, not a feature built into Health-Nut AI itself. Any text box in the app (a search field, a recipe request, a follow-up question) can be dictated: tap the microphone icon your phone's own keyboard already shows next to the space bar (iPhone and Android both have one), speak, and it types what you said. Handy any time typing is inconvenient, like standing in a grocery aisle.