How I Lost 12 Pounds in 12 Weeks: Building and Using My Own Nutrition Tracking App
In early May 2025, I stepped on the scale and saw 165 pounds staring back at me. It wasn't a shocking number, but combined with my recent blood work showing elevated glucose and HbA1c levels (6.4 - prediabetes territory), it was a serious wake-up call.
I had been consistently doing 3-4 workouts per week for nearly two years - running, swimming, cycling - but my weight had remained stable around 165 lbs. Despite being active, my metabolic health wasn't where it needed to be. During my follow-up appointment, my primary physician was direct: "You need to lose 10 pounds. Exercise is great, but nutrition is what will move the needle on your metabolic health."
I realized that exercise alone wasn't enough; I needed to pay attention to nutrition.
That's when I decided to build a nutrition tracking app that would actually work for me - one that used AI to make logging effortless and provided intelligent insights. That app became Macrosia. Over the next 12 weeks, I tracked every meal while building and refining the product. I dropped from 165 to 157 pounds - a steady 8-pound loss that taught me invaluable lessons about sustainable weight management and product development.
The Decision to Get Serious
In early May 2025, with prediabetes knocking at my door despite years of regular exercise, I realized I needed a systematic approach to nutrition tracking. Rather than use existing apps that felt clunky and disconnected from my workflow as a developer, I decided to build my own solution.
I made a dual commitment: build a nutrition tracking app that actually worked for me, and use it to track everything I ate for the next three months. This would let me create and test the product simultaneously while addressing my health concerns.
The Data Behind the Journey
Before diving into the story, let me show you exactly what happened with my weight over those three months:
What's remarkable about this chart is the consistency. No dramatic drops, no yo-yo effects - just steady, sustainable progress at about 1-1.5 pounds per week.
Setting Up My Goals
As I built out Macrosia's profile setup feature, I entered my own stats as the first test case:
- Height: 5'7"
- Starting weight: 165 lbs
- Goal: Gradual weight loss (1-2 lbs per week)
- Activity level: Moderately active (3-4 workouts per week)
I programmed the app to calculate daily targets based on established nutrition science:
- Calories: 1,800-2,000
- Protein: 140-160g
- Carbs: 180-220g
- Fat: 60-80g
One principle I built into Macrosia from day one was avoiding impossibly restrictive targets. I designed it to provide ranges that felt achievable and sustainable, which proved crucial for my own adherence.
The Nutrition Reality Check
Here's what my daily nutrition actually looked like over the journey:
Several key insights jump out from this data:
- Calorie consistency: I averaged 1,728 calories per day - right in the sweet spot for gradual weight loss
- Protein optimization: My protein intake actually increased over time, averaging about 140g daily
- Carb moderation: I naturally gravitated toward fewer carbs as I got more active
Monthly Breakdown: What Actually Happened
Month 1 (May): Building Habits (-4 lbs)
Starting: 165 lbs → Ending: 161 lbs
May was all about establishing the tracking habit. I was religious about logging but didn't change my eating patterns dramatically.
Key discoveries:
- I was eating 200-300 more calories than I thought
- My protein was consistently 20-30g short of targets
- Weekend eating was significantly higher than weekdays
What worked:
- Morning meal planning during coffee
- Natural language logging I built in ("grilled chicken salad with balsamic")
- Early AI Coach prototype I developed after collecting 3-4 days of data
Month 2 (June): Finding My Rhythm (-4 lbs)
Starting: 161 lbs → Ending: 157 lbs
By June, logging was automatic. More importantly, I was starting to see patterns I'd never noticed before.
Pattern recognition:
- Weekend overeating: 300-400 calories higher than weekdays
- Stress eating: High-carb, low-protein days during busy work periods
- Exercise timing: Morning workouts = better food choices all day
As I developed the AI Coach feature during this period, it became incredibly valuable at recognizing these patterns and offering proactive suggestions:
"Your weekend intake has been higher lately. Consider planning weekend meals in advance or having healthy snacks readily available."
But the real game-changer was when I built AI Coach insights about workout nutrition timing. After feeding it my eating patterns and workout schedule, I programmed it to make several key recommendations:
- Carb timing: Concentrate carbs around workouts (30-60g before, 30-40g after)
- Protein increase: Boost daily protein from ~100g to 140-150g to support my 4-5 weekly workouts
- Fat reduction: Lower fat intake slightly to make room for more protein without exceeding calories
- Fiber optimization: Increase fiber to 30-35g daily for better satiety and digestion
These weren't generic recommendations—they were based on my specific workout schedule, meal timing, and body composition goals.
Month 3 (July): Optimization and Plateau (-1 lb)
Starting: 157 lbs → Ending: 157 lbs
July was when both the app and I moved from basic tracking to truly understanding my body's needs.
The plateau breakthrough: Mid-July, I hit a 10-day plateau at 156.5 lbs. Instead of getting frustrated, I built and used Macrosia's trend analysis feature. The data showed consistent calories but declining protein and inconsistent meal timing due to travel.
Adjusting these two factors broke the plateau within 3 days.
The Workout Connection
Here's how my exercise routine supported the nutrition changes:
Exercise insights from the data:
- Consistency over intensity: I averaged 4-5 workouts per week
- Mix of activities: Running (2x/week), swimming (1-2x/week), cycling (1-2x/week)
- Smart nutrient timing: Following AI Coach guidance, I timed carbs around workouts and increased post-workout protein
- No over-compensation: Exercise didn't lead to significantly higher calories, thanks to strategic meal timing
My Macro Evolution
Here's how my macro balance shifted over the three months:
What this reveals:
- Protein priority: I consistently hit 140-150g daily (vs. my previous ~100g)
- Moderate carbs: Averaged 158g (perfect for my activity level)
- Healthy fats: 67g daily from nuts, avocado, olive oil
This wasn't a "diet" - it was learning to eat in alignment with my body's actual needs.
What Made This Different
I've tried losing weight before, but this was different. Here's why:
1. No Judgment, Just Data
Traditional calorie counting feels punitive. Macrosia's AI Coach felt like having a supportive nutritionist who understood my patterns and offered practical solutions.
2. Natural Language Input
Being able to type "leftover pizza, 2 slices" instead of searching for "Pizza, cheese, large slice, Pizza Hut" removed all friction from logging.
3. Intelligent Insights
The AI didn't just track—it taught me. I learned about nutrient timing, satiety, and how different macros affected my energy and performance.
4. Flexibility Within Structure
Having ranges instead of rigid targets meant I could adapt to real life while staying on track.
The most powerful insight: On days when I hit my protein target early, I naturally ate fewer total calories without feeling restricted. This taught me that protein timing is everything for appetite control.
The Unexpected Benefits
Beyond the 8-pound loss, building and using Macrosia gave me:
- Better sleep: Consistent meal timing improved my circadian rhythm
- Stable energy: No more afternoon crashes
- Improved performance: My workouts became more productive with optimized pre/post nutrition
- Food confidence: I can now estimate portions and macros accurately without logging
- Improved metabolic health: My HbA1c dropped from 6.4 to 5.9, moving me out of prediabetes range
What I'd Tell Someone Starting Today
-
Start with tracking only: Don't change your eating habits for the first week. Just log everything and let the AI show you patterns.
-
Trust the ranges: Don't aim for the exact middle of every macro target. The ranges exist for a reason.
-
Use the AI Coach consistently: Generate insights every few days. It gets better at understanding your patterns over time.
-
Focus on trends, not daily fluctuations: Weight varies day to day. The weekly trend is what matters.
-
Be honest about everything: The app can only help if it has accurate data.
-
Prioritize protein: This was my biggest game-changer for appetite control and body composition.
The Real Success: Sustainable Habits
The 8-pound loss is great, but the real success is that I now have an intuitive understanding of nutrition that will last forever. I know what 30g of protein looks like. I know how different carb sources affect my energy. I know when I need to eat more and when I need to eat less.
Most importantly, my physician is thrilled with the results. My HbA1c improvement from 6.4 to 5.9 represents a significant step back from prediabetes, and my blood glucose levels have stabilized. When he saw these latest results, he told me: "Keep doing exactly what you're doing, and let's aim for another 10 pounds." This validation from a medical professional gave me confidence that my approach is not only safe but sustainable for long-term health.
I also proved to myself that the app we built actually works—not just for tracking, but for creating lasting behavior change that a doctor can endorse.
The Data Doesn't Lie
After analyzing 1,000+ meal entries and 65 workouts, the formula for my success was surprisingly simple:
- Consistent calorie deficit: 200-300 calories below maintenance
- Protein prioritization: 140-150g daily
- Regular exercise: 4-5 sessions per week
- Flexible consistency: 80/20 rule, not perfection
- Data-driven adjustments: Let the trends guide changes, not daily fluctuations
Want to start your own journey? Try Macrosia free and see what insights your data reveals. The first week of discovery might surprise you as much as it did me.