Early Access — iOS Beta

Snap your meal.
Your log is done.

For gym-goers who've already tried calorie tracking and quit — this logs a meal from a photo in under 30 seconds, so the habit finally sticks.

Built for adults 18–40 who lift 3–5×/week and know exactly what they want: a number goal and a plan that doesn't waste their time.

Your day, at a glance

Tap any meal card to expand macros. Tap "Log a Meal" to see the snap flow.

9:41
snap&track
🏋️
1,480
kcal logged today
▲ 620 remaining toward 2,100 goal
118g Protein
142g Carbs
44g Fat
Breakfast · 7:42 am
🥣
Greek Yogurt + Granola + Berries
Snapped · 7:44 am · 98% confident
380 kcal
Identified from photo
32g
Protein
48g
Carbs
9g
Fat
Confidence
98%
Portion: 1 bowl (~320g) · Tap to edit
Lunch · 12:18 pm
🍗
Chicken Rice Bowl w/ Broccoli
Snapped · 12:20 pm · 91% confident
610 kcal
Identified from photo
52g
Protein
64g
Carbs
14g
Fat
Confidence
91%
Portion: large bowl (~480g) · Tap to edit
Snack · 3:30 pm
🥤
Protein Shake + Banana
Snapped · 3:32 pm · 95% confident
490 kcal
Identified from photo
34g
Protein
30g
Carbs
21g
Fat
Confidence
95%
Portion: 1 shake + 1 banana · Tap to edit
📸
Dinner not logged yet
Tap "Log a Meal" below to snap it
🍝
Point at your meal
The app will identify it automatically.
No barcode, no database search.
Identifying your meal…
~3 seconds

Tap any meal card to expand macros · Tap "Log a Meal" to see the snap flow

Three steps, under a minute

Step 1
Snap the photo

Open the app, point your phone at any meal — home-cooked, restaurant, or packaged — and tap once. No setup.

Step 2
Macros log instantly

Calories, protein, carbs, and fat appear in under 30 seconds with a confidence score. Tap to adjust portion size if needed.

Step 3
Your plan adapts

The app watches your real logged adherence week over week and adjusts your meal plan and workout split automatically — no manual reconfiguring.

⚠ Sample testimonials — pre-launch. No paying users yet.

"I've started and quit MFP three times. The search bar alone was enough to make me give up after two weeks. If snapping actually works accurately, I'm in permanently."

💪
Marcus, 28
Lifts 4×/week · Goal: +10 lbs muscle

"Restaurant meals always killed my tracking. I could never find the right item in the database and would just give up. A photo would actually solve that."

🥗
Priya, 33
Runs + lifts · Goal: −15 lbs in 12 weeks

"My problem isn't knowing what to eat — it's the 5 minutes of data entry per meal. That's what makes tracking feel like a second job. Cut that and I'll actually stick to it."

🏃
Derek, 24
CrossFit 5×/week · Former Lose It! user

Things you should ask before joining

That's the most important question to ask, and we want to be straight with you: Lose It! does have photo recognition, and it's backed by a large engineering team. The honest answer is that the competitive bet here is not "we have photo logging and they don't" — it's that a focused app built around this single mechanic, with no legacy UI debt, can iterate faster on recognition accuracy and make the experience meaningfully faster end-to-end. Whether that's true is something we're actively testing. That's why we're building with a beta cohort before claiming any advantage. If accuracy testing shows we can't clear 85%+ on unstructured home-cooked meals, we'll tell you — and we'll share those results.

We won't sugarcoat this. The brief we wrote for ourselves sets the minimum bar at ~85% accuracy on unstructured home-cooked meals. Field-wide accuracy today (across available models) is estimated at 60–75% for that meal type — which means we may not be there yet at launch. For structured meals (protein + starch + vegetable bowls, packaged foods, common restaurant dishes), accuracy is substantially higher. Our kill criteria include a blind test of 50 user-submitted home-cooked photos: if we fall below 80%, we don't launch broadly. When the model is unsure, the confidence score will show it — and we'll prompt you to confirm rather than silently log a wrong number.

This is the hardest question, and we don't have a complete answer yet. Here's what we believe and what we're testing: faster logging reduces the daily friction tax, but it probably isn't sufficient on its own. The adaptive plan — one that changes based on what you actually ate rather than what you planned to eat — is our hypothesis for the week-4 mechanic. If the app responds to your real behavior rather than guilt-tripping you for a missed day, re-engagement is easier. We're measuring day-14 retention in the beta cohort specifically. Our threshold is 25%+. If we can't hit it, we'll tell you what we found before opening up more broadly. The research is honest: this is the central unsolved problem in the category.

Help us decide what to build first.

We're testing with a small iOS beta cohort. Your answer below tells us whether we're solving the right problem — and shapes what ships in v1.

We read every reply.