The ones people ask before they trust a number. Where the honest answer is that the data cannot tell you, that is the answer you get.
What it is, where it comes from, and what it cannot see.
Statistically it is the eating, and that is not an accusation. Under-reporting is unconscious and close to universal. But you do not have to settle this with population statistics, because your own numbers can answer it directly.
On the metabolism half of the question: between two people of the same size, sex and body composition, the honest spread is a few hundred calories a day, not the thousand people imagine. Most people who feel like an outlier are not one.
After enough complete days the app puts your measured burn next to what the formula predicts for a body your size and names the gap in calories. If you genuinely sit below prediction, it says so plainly. If you do not, the same measurement points you at the deficit arithmetic instead.
Eating more does not create a deficit. That part is arithmetic, and it holds however slow anyone's metabolism is.
What is real, and worth naming rather than laughing at: under-reporting rises sharply while someone is dieting, unconscious daily movement drops when energy is low, and very low intakes wreck adherence within weeks. A stall is usually one of those three, not a switch that flipped inside you.
You do not have to take my word for it. Your own measured expenditure over the next few weeks tests the claim directly, and if you genuinely are burning far less than your size predicts, that is what the number will say.
Neither the frightening yes nor the dismissive no.
Adaptation is real and, for most people, modest. After weight loss people tend to burn somewhat less than body size predicts, on the order of a couple of hundred calories a day, and it scales with how much was lost and how fast it came off.
The televised contestant data is an extreme, not a template. Those were extraordinary deficits, extraordinary training loads and extraordinary rates of loss, and very few people have done anything comparable.
There is no protocol here that claims to undo it, because nothing sold under that banner has controlled evidence behind it. What does have evidence is a stretch of eating at maintenance: expenditure tends to return toward what size predicts. That is what it does, and that is all it does.
Not in the way it is usually described. Pooled doubly labeled water data covering thousands of people shows size-adjusted expenditure holding roughly steady from the twenties to about sixty, with no distinct step at menopause in that dataset.
What does change over those decades is body composition, activity and sleep, and those change what you burn.
This is population data. It does not overrule an individual endocrine situation, and it is no reason to ignore one. What it means here is that TrueBurn applies no age penalty triggered by a birthday. It measures what is actually happening to you.
A calculator predicts. This measures.
A TDEE calculator takes your height, weight, age, sex and an activity band you picked for yourself, then returns what someone with those inputs tends to burn. That band is a guess you made about yourself, and it is usually the largest single error in the answer.
TrueBurn runs it the other way round. What you ate across a window, plus what your weight trend actually did, implies what you must have burned:
The limits are real and worth saying in the same breath. It needs about two weeks of complete days before it says anything, and it inherits whatever bias is in your food log.
Yes. That is worth conceding first, because it is true.
Average under-reporting runs around 12 to 16 percent, and higher while someone is actively dieting. If you log 15 percent low every day, your measured expenditure reads about 15 percent low.
It is still useful, because the target it produces is expressed in the same units you log in. Log 15 percent low, get a target 15 percent low, and you still land at the intended rate of loss.
What breaks it is inconsistency, not inaccuracy. Getting more careful halfway through a window moves the number for no physiological reason. Log the same way on good days and bad days, and distrust any reading taken across a stretch where your habits changed.
About two weeks of complete days. Before that you see a formula estimate, labeled as an estimate, never dressed up as a measurement.
The window exists because a weight trend needs enough days for daily water swings to average out against the signal underneath them. Four days of scale readings are mostly water.
It does not turn from wrong to right on a particular morning. It gets less shaky. The confidence indicator comes from the sample count and calendar span of the window actually in use, not from how long you have had the app.
They measure different things, so more accurate is the wrong axis.
A clinic test measures your resting rate over about ten minutes, fasted and still. It is a precise snapshot of one component, and it still does not tell you your total daily burn.
TrueBurn measures total daily expenditure, including everything you actually did, across weeks, for free, at lower precision and with your logging bias sitting inside it.
Book the lab test if you want your resting rate. This answers a different question.
No, and logging them would count the same energy twice.
The measured figure comes from your intake and your weight change, so everything your body spent is already inside it: training, walking, standing, fidgeting. There is nothing left over to add back.
It is also why wearable estimates are not an input. Tested against reference methods, their calorie numbers have come out off by roughly 27 to 93 percent.
A genuine change in training volume does show up. Add three sessions a week and the measured number moves on its own within a couple of weeks, and you can watch it happen.
Three pounds of fat is more than ten thousand surplus calories. You did not eat ten thousand surplus calories yesterday.
What moved is water. Each gram of stored carbohydrate holds roughly three grams of water with it, so a big meal, more salt than usual or a hard session can put a few pounds on the scale overnight and take them off again just as fast.
This is why the app runs off a smoothed trend rather than the morning reading. Both lines are drawn together, so you can check that claim instead of being asked to believe it.
Because two weeks of data cannot separate the candidates, and an app that picks one anyway has invented it.
A stall fits several explanations at once: logging drift, a real drop in expenditure, water retention hiding fat loss that is still happening, or a window too short for the noise in it. Over a fortnight those look nearly identical on the chart.
What separates them is more evidence. A longer window. A waist measurement that keeps falling while the scale does not. A careful stretch of logging compared against a normal one.
Pointing at which explanation your own data leans toward is planned work and is not built. Today the app shows the candidates and the evidence and leaves the conclusion with you.
Who pays, where your log lives, and how you lose it.
People say that if the product is free, you are the product. That is not what is happening here, and the rest of this answer is why.
No catch. I built this for myself, it helped, so I put it out for free in case it helps someone else. There is no revenue, no advertising and no data deal.
There is also nothing here to sell. Your logs never leave the device, so there is no dataset to license. You can check that yourself: turn the network off after the first load and the app keeps working.
A paid sync tier is planned, for people who want their history on more than one device. That is a different product with real server costs. If any of the above ever changes, it changes in the changelog and on this page before it ships, not quietly.
If it is useful to you and you want to send something, there is a donate link. It is entirely optional and nothing about the app changes if you use it.
This is a permanent commitment, not a launch promise. Private local-only mode stays free. Paid tiers, when they exist, will be for people who want their data synced across devices, and that is a genuinely different product with genuinely different costs. Nobody will ever have to hand over their weight history to keep using what they already have.
In your browser, in IndexedDB, on the device you logged it on. No, I cannot see it, because there is no server holding it.
One thing does leave the device: the food database lookup when you search for an item. It sends the search text, it carries no account and nothing that identifies you, and it is the only call that carries anything you entered.
You can check this in about thirty seconds. Open your browser's developer tools, watch the network panel while you log a weigh-in and a meal, and see exactly which requests are made.
You lose the log. Clearing site data deletes it, a private window keeps nothing once you close it, and browsers can evict storage on their own when a site has not been opened for a while. iOS is the strictest about that.
Your history exists in exactly one place with no backup. That is the cost of keeping it off a server, and it is the reason export exists.
So: the export button is never more than one screen away. Use it whenever losing a month of logging would annoy you, and keep the file somewhere you already back up.
Not together, in this version. Local storage means each browser holds its own separate history, so your phone and your laptop are two different logs.
The workaround is manual: export from one, import into the other. It is a snapshot rather than a merge, so if you keep logging on both they drift out of agreement.
Sync is planned. It needs accounts and a server, which is the line this version deliberately does not cross, so it would arrive as an opt-in paid tier. No date.
Your history is not on a server, so there is nothing anyone can switch off that takes it with them.
Your export is a plain readable file you already hold, in a documented format that other software could read. Every schema change ships with a forward migration, every read tolerates records written by older versions, and if the format ever changes the import keeps reading the old one.
That is the whole answer. The part of this project that sounds smallest is the part that outlasts me.
Who this is not for, and what it cannot diagnose.
Here is what the evidence shows, rather than a reassurance. Short controlled trials in general populations have not found harm from calorie tracking. People with existing risk report that it can amplify what is already there, and the concerning pattern is tracking plus predisposition plus rigid, all or nothing restraint.
If that describes you, talk to a professional before you start. That is the actual advice, not a disclaimer at the bottom of it.
What the app does instead of calling itself safe: no leaderboards, no friend feeds, no ranking, no "you are in the top 20% of users". Progress is always you against your own last month.
Some of that work has since been done. The streak reads as a plain count of days rather than a flame you can lose, nothing is scored during a precision week, and the daily cards were softened. This page will keep saying what is still outstanding rather than claiming the question is closed.
It does not model your condition, and it does not need to, because it measures the outcome.
Whatever a condition or a medication is doing to what you burn is already inside your weight trend and your intake, so it lands in the measured number without anyone having to name it.
What it cannot do: tell you why the number is what it is, identify a condition, or stand in for your clinician. It is not medical advice.
One thing worth doing anyway. If your measured number moves sharply and nothing about your eating explains it, that is worth mentioning at your next appointment.
If I cannot explain how I got a number, I do not show the number.
The long version of the argument, including where other apps beat mine, is on the why page. Or start logging and see what your own two weeks say.
Free. Private. No sign-up.