Show HN: Lean bulk, cut, body recomp. Calculate maintenance calories(macrocodex.app)
macrocodex.app
Show HN: Lean bulk, cut, body recomp. Calculate maintenance calories
https://macrocodex.app/
12 comments
> when you eat more than your maintenance calories, you gain weight; when you eat less than your maintenance calories, you lose weight.
Sounds interesting.
At it's core, this idea does roughly work. But there are a lot of caveats around maintenance calories.
The main one is that not all calories are equal. 500 calories of spinach is very different to 500 calories of a McDonalds burger.
Is this something that the service is able to take in to account?
Sounds interesting.
At it's core, this idea does roughly work. But there are a lot of caveats around maintenance calories.
The main one is that not all calories are equal. 500 calories of spinach is very different to 500 calories of a McDonalds burger.
Is this something that the service is able to take in to account?
>The main one is that not all calories are equal. 500 calories of spinach is very different to 500 calories of a McDonalds burger.
It doesn’t ask what specific foods you eat, so we do not have this data.
Interestingly, that isn’t a problem. Most people follow fairly predictable eating patterns over time. We measure the effectiveness of our method by comparing the user’s target rate with the actual rate of loss or gain achieved whether that’s weight loss or weight gain.
Our data tells us 56% of users start losing weight or gaining weight (depending on set goal) within 2 weeks of using macrocodex and 95% start seeing meaningful gain and loss trend within 5 weeks.
Based on this alone, we'll probably never add more granular data collection than this.
It doesn’t ask what specific foods you eat, so we do not have this data.
Interestingly, that isn’t a problem. Most people follow fairly predictable eating patterns over time. We measure the effectiveness of our method by comparing the user’s target rate with the actual rate of loss or gain achieved whether that’s weight loss or weight gain.
Our data tells us 56% of users start losing weight or gaining weight (depending on set goal) within 2 weeks of using macrocodex and 95% start seeing meaningful gain and loss trend within 5 weeks.
Based on this alone, we'll probably never add more granular data collection than this.
It's not only that. Also, absorption of calories from food is largely different between individuals, due to gut microbiome, generics, etc.
I like this, I’ve been tracking with MyFitnessPal. Would be great if I could import historical calories to see how it performs.
Also note that tracking calories is hard - often it’s an estimate so quite hard to vary by the small amounts above and below the threshold needed for continued weight loss.
Also losing weight is hard!
On Android, it supports Google Health Connect.
Users can import up to 30 days of calorie and weight data from other apps that write to Google Health Connect.
Also, yes, Macrocodex doesn’t change your targets for every minor fluctuation in maintenance calories. It only adjusts your target when enough meaningful and measurable change has accumulated, as you correctly pointed out tracking changes below ~150 kcal is difficult.
Users can import up to 30 days of calorie and weight data from other apps that write to Google Health Connect.
Also, yes, Macrocodex doesn’t change your targets for every minor fluctuation in maintenance calories. It only adjusts your target when enough meaningful and measurable change has accumulated, as you correctly pointed out tracking changes below ~150 kcal is difficult.
Do i have to put in what I ate throughout the day? How accurate / easy is it for me to put in each ingredient that I ate and its weight? Instead, can i put in a weekly meal plan that I already have?
No, this isn't a calorie tracking app.
You need to log only two things
1. Total calories consumed on a specific day
2. Body Weight
it doesn't care about anything else.
You need to log only two things
1. Total calories consumed on a specific day
2. Body Weight
it doesn't care about anything else.
It’s a great product. I’d mainly like to know: how does this compare to connecting Apple Health data to Claude or the ChatGPT app and having them perform calculations in a specific way?
Macrocodex can work even when calorie logging is sparse. A user may fail to log data for a few days each week, and maintenance calories can still remain accurate because it operates over your entire history.
If you read our privacy policy, you’ll see that it collects total calorie and weight data from users, which we use to improve our adaptive TDEE algorithm. It does not collect email, phone number, date of birth or anything which is not necessary for product improvement/diagnosis.
You can probably achieve something similar with Claude and a bunch of scripts if you log your calories every day.
macrocodex is entirely deterministic, it just works and users don't need to fiddle with it.
Log your calorie each day, log your weight weekly and it will continue producing calorie and macro targets for you to achieve desired rate of weight gain or loss.
It also has many more features, such as lean bulk, cut, and body recomposition modules. It includes a planner that helps you decide when to switch from cutting to lean bulking (or vice versa), or when body recomposition efficiency starts dropping for you.
For example,
To gain least fat while gainng muscles, Lean bulk is what you need.
The key is to focus on your rate of weight gain, not a specific calorie number or surplus percentage.
Weekly Weight Gain Rate:
Beginners: 0.25-0.5% of body weight
Intermediates: 0.25-0.4%
Advanced Lifters: 0.1-0.25%
Usually, a calorie Surplus: +5-10% of TDEE (roughly 100-300 kcal/day) may get you there but if it doesn't adjust your calories up or down to reach the desired weekly weight gain rate.
This is where macrocodex can do better than chatgpt/claude over long horizon.
If you read our privacy policy, you’ll see that it collects total calorie and weight data from users, which we use to improve our adaptive TDEE algorithm. It does not collect email, phone number, date of birth or anything which is not necessary for product improvement/diagnosis.
You can probably achieve something similar with Claude and a bunch of scripts if you log your calories every day.
macrocodex is entirely deterministic, it just works and users don't need to fiddle with it.
Log your calorie each day, log your weight weekly and it will continue producing calorie and macro targets for you to achieve desired rate of weight gain or loss.
It also has many more features, such as lean bulk, cut, and body recomposition modules. It includes a planner that helps you decide when to switch from cutting to lean bulking (or vice versa), or when body recomposition efficiency starts dropping for you.
For example,
To gain least fat while gainng muscles, Lean bulk is what you need.
The key is to focus on your rate of weight gain, not a specific calorie number or surplus percentage.
Weekly Weight Gain Rate:
Beginners: 0.25-0.5% of body weight
Intermediates: 0.25-0.4%
Advanced Lifters: 0.1-0.25%
Usually, a calorie Surplus: +5-10% of TDEE (roughly 100-300 kcal/day) may get you there but if it doesn't adjust your calories up or down to reach the desired weekly weight gain rate.
This is where macrocodex can do better than chatgpt/claude over long horizon.
That sounds great, but it feels inappropriate to collect users' calorie data without explicitly informing them.
Calorie and weight data collected by the Android app looks like this:
`datetime | 2500 kcal | 90 kg`
By using Macrocodex, users become part of a shared data pool. They pay nothing, but they can choose to share this data.
We've this data from 17,000+ users that's what makes this accurate.
Data collection on Android can be disabled by turning off Diagnostics under Profile.
The app does not collect anything beyond this, so there is no way to link the data back to any specific person. It does not even ask for your name.
Furthermore, users can easily delete all collected data by going to *Profile > Clear Data*. During onboarding, the app will also show the Health Connect permission request.
On the web app, there is no reliable way to store persistent data because IndexedDB storage can be wiped by the OS or browser. Because of this, it requires backend to save this data.
There are no cookies used on site, no tracking pixels etc...
`datetime | 2500 kcal | 90 kg`
By using Macrocodex, users become part of a shared data pool. They pay nothing, but they can choose to share this data.
We've this data from 17,000+ users that's what makes this accurate.
Data collection on Android can be disabled by turning off Diagnostics under Profile.
The app does not collect anything beyond this, so there is no way to link the data back to any specific person. It does not even ask for your name.
Furthermore, users can easily delete all collected data by going to *Profile > Clear Data*. During onboarding, the app will also show the Health Connect permission request.
On the web app, there is no reliable way to store persistent data because IndexedDB storage can be wiped by the OS or browser. Because of this, it requires backend to save this data.
There are no cookies used on site, no tracking pixels etc...
This is very helpful to me.
By using an algorithm, we can accurately figure out your maintenance calories more accurately than traditional regression based formulas like katch mc ardle.
It's way more accurate than calorie burn tracking devices like fitness bands and watches. (garmin/apple watch/whoop etc...)
Traditionally, people often use static TDEE calculators which often over or underestimate for some by 100s calories.
Chatgpt and TDEE calculators like Calculator.net or TDEECalculator.net use the same formulas, so they share the same limitation
If a beginner asks ChatGPT, "What are my maintenance calories?", ChatGPT can give them a number. But ask how it arrived at that number, and it will usually explain that it used a formula like Katch McArdle, Harris Benedict, or Mifflin St Jeor to calculate BMR, then layered activity on top using an activity factor, PAL, or MET tables.
Dig deeper and those formulas come from statistical regressions based on averages from past populations. That means maintenance calories calculated this way can be off by hundreds of calories.
MacroCodex uses your calorie intake and weight data to figure out maintenance calories specific to your body, not the population average.
It usually reaches good accuracy after about 3 weeks, or 21 days of calorie and weight logging.
This app is completely free, no paywall, no subscription and no ads. (works offline)
Most people start seeing weight gain or loss within 5 weeks.
We've reached 13,000+ users. Full support is provided if you experience any issue, most issues are resolved under 24 hours.