Snap it. Say it.
Scan it. Traq it.

AI nutrition tracking for iPhone. Snap a photo, speak a sentence, or scan a barcode — and when the AI gets something wrong, just tell it.

See how it works ↓

Free during beta · iOS 18.0 or later · iPhone only

How it works

Three ways in. One source of truth.

Most trackers offer a search box. We don't. You log a meal the way it actually happens — by photo, by sentence, or by scan — and the rest is the app's problem.

Snap a photo. Get full nutrition.

Identifies every item on your plate, estimates portions, and gives you the full nutrition breakdown. From a single photo.

A meal photo analyzed by the app, with each food item outlined and labeled — salmon fillets, sautéed spinach, mixed vegetable salad, and seed-crusted bread rolls — alongside total calories and macros.

Just say what you ate.

A conversational assistant understands plain language. Type or speak — "I had eggs and toast", "yesterday at 8pm a slice of pizza" — and the meal lands on the right day, on the right plate, with the right macros. The mic in the chat works hands-free for when you're cooking.

The chat assistant rendering a Markdown summary of yesterday's meals with bulleted breakdowns and calorie totals, overlaid on the day's meal log.

Scan a barcode. Instant lookup.

Point the camera at any barcode and get a full nutrition card immediately. No typing, no searching, no second-guessing the database row.

Scanning a yogurt-cup barcode in the app, with a nutrition card surfacing showing calories, protein, carbs, and fat per serving.

There's also a one-tap water button — long-press for other sizes — with an opt-in reminder if you're falling behind on your hydration goal.

When the AI gets it wrong

Don't delete it. Talk to it.

Every other photo tracker leaves you stuck when the AI mislabels an item — delete the meal, log it again from scratch. Not here. Tap any item in your meal and you're in a conversation about that specific item. The assistant already knows what it currently thinks the item is, and what the photo of it looks like. Tell it what's wrong in plain language:

  • "That's a chicken wrap, not a tortilla." — the AI re-analyzes that item and updates it.
  • "Make it 200 g, plus a tablespoon of olive oil." — direct edit, macros recalculate.
  • "Rename it to 'lunch wrap'." — done.

Chain corrections in one conversation. The macros, the photo annotation, the calorie total — everything updates as you go. And there are quick shortcuts for the common fixes: swap an item, add one the AI missed, or redraw the region you want re-analyzed.

The chat assistant scoped to a single food item from a meal, with the cropped photo of that item pinned above the conversation, and the user telling the AI what the item actually is so it can re-analyze.

Accuracy

The AI does the boring part. Then checks itself.

Most photo-logging apps run a single classifier and call it done. We don't. Our pipeline is multi-stage and self-checking — when the calorie estimate looks off against the photo, the AI re-examines the image before committing the result. And when you spot something it missed, correcting it takes one sentence.

Meal detail screen showing a multi-item plate analyzed by the AI: each detected item is outlined and labeled on the photo, and listed below with its own calories and macros.

Goals & energy

Targets you can actually express.

Six nutrients, three directions each

Calories, protein, carbs, fat, fiber, water — each one can be at least, at most, or a target. So fiber can be "at least 30 g" while calories is "at most 2000". Protein supports per-kg body-weight scaling, so it stays right when your weight changes.

Energy balance, in plain numbers

Today's net — what you ate minus what you burned — and a target you set yourself. Whether you're cutting, bulking, or just maintaining, the app shows you whether you're tracking against it.

Apple Health, both directions

Reads your active and resting energy expenditure straight from HealthKit; writes the nutrition you log back to it. One source of truth across the Apple ecosystem, no manual entry on either side.

Outside the app

Your fasting timer, one glance away.

Intermittent fasting is built in — pick a ratio (14:10, 16:8, 18:6, 20:4) or set a feeding window. Lock Screen and Home Screen widgets keep the timer one glance away, no app launch required. The widget turns green the moment you hit your fasting goal.

A home screen widget showing the traq fasting timer — a circular gauge with elapsed time toward the fasting goal.

Privacy

Your data, never anyone else's product.

Photo analysis and chat run on trusted cloud AI providers (AWS, Alibaba Cloud), used only to recognize your food — those providers don't train their own models on your data. We may use beta usage data to improve traq's own food-recognition accuracy. Your data is never sold, never shared with advertisers or data brokers, and never used for ad targeting. Sign out and the device forgets you locally; delete your account and the server forgets you too, after a 30-day cancellation window.

Read our privacy policy →

Built in

And more.

  • Calendar with goal dots

    Every day in the calendar shows a status dot per goal so you can see your adherence at a glance and drill into any single day.

  • Smart push notifications

    Opt-in meal reminders that aren't spam, plus a ping the moment a photo finishes analyzing so you can tap straight to the result.

  • Hands-free voice input

    The mic in the chat works while you cook. Speak, log, move on.

  • Works offline

    Photos queue locally and process the moment you reconnect.

Start tracking in seconds.

traq.food is in private beta on TestFlight. Join the list and we'll send you access. Pricing for the public release will be announced before it ships — no surprise paywalls.

Free during beta · iOS 18.0 or later · iPhone only