A free calorie counter app with no subscription: what to expect

If you are searching for a free calorie counter app with no subscription, here is the honest answer: genuinely free calorie logging exists and always will, but only for the parts of the job that cost the developer nothing each time you use them. Adding up numbers, storing text, applying a formula, drawing a trend line: these are free to run, so an app can give them away forever without lying to you. Recognising a photo of your dinner is not free to run, and no app can offer that without limit unless somebody else is paying. Once you can tell those two categories apart, you can predict which free tiers will still be there in a year.

I build a nutrition app, so I have had to make these decisions with real numbers in front of me. What follows is how I would evaluate a free tier if I were the one installing it.

The marginal cost line

Here is the distinction I use, and I have not seen anyone state it plainly, so I will name it.

The marginal cost line is the boundary between features that cost the developer nothing per use and features that cost something every single time they run. Above the line, free is sustainable indefinitely. Below it, free is a promotion, a subsidy, or a limit you have not hit yet.

Above the line:

• Calculating a calorie target from your height, weight, age, sex and activity level. This is arithmetic.
• Splitting that target into protein, carbohydrate and fat.
• Storing a text record of what you ate and adding it up.
• Computing a seven-day weight trend from daily entries.
• Any score or streak or summary produced by a stated rule rather than a model.
• A food database lookup. Serving a row from a table is close enough to free at consumer scale.

Below the line:

• Vision analysis of a meal photo. Every photo is a separate inference that costs money.
• Text generated by a language model, including chat, coaching and written insights.
• Voice transcription.
• Barcode scanning against a commercially licensed database, where the licence is priced per query.

This is why the pricing pages of nutrition apps look the way they do. It is not arbitrary. Draw the line yourself, and a free tier stops being a mystery.

What this tells you about any free calorie counter

Two useful predictions come out of the marginal cost line.

First: if an app gates the arithmetic, it is gating arithmetic. Charging a subscription to see your macro split, or to change your activity level, or to view a weight trend, is not covering a cost. It is a decision to hold back something cheap because it is the thing you will miss. That is legitimate business, but you should recognise it for what it is rather than assuming the price reflects difficulty.

Second: unlimited AI photo recognition, free forever, is not a stable offer. If you see it, one of three things is true. Someone is subsidising it to buy growth, the model being used is small enough to be cheap and the accuracy shows it, or the terms will tighten. I would put a specific number on it: an app offering unlimited free vision-based meal logging today will have introduced a cap, a queue, ads, or a paywall within about a year. That is a falsifiable claim and I am happy to be held to it.

None of this means avoid free tiers. It means read them as a forecast rather than a promise.

The retroactive paywall, and why it is the only red flag that matters

Gating a new feature is fine. Gating the past is not.

I call it the retroactive paywall: when an app restricts access to data you already entered. You log conscientiously for six weeks, then the app decides that history beyond the last fourteen days is a premium feature. The work was yours. The typing was yours. The paywall arrived afterwards.

This is the single thing I would check before investing effort in any free calorie counter, because logging has a switching cost that grows every day you keep at it. The check takes two minutes:

1. Look for an export. Any format will do. If you can get a file out, you are never trapped.
2. Search the app's own description or help text for a history limit. Phrases like "last 7 days" or "recent entries" in a feature comparison are the tell.
3. Check whether the app works offline for viewing. If your own past requires a server round trip, your access depends on a business decision.

An app that will show you everything you ever entered, forever, at no cost, has made a commitment that is expensive to reverse. That is worth more than any feature list.

Work out your own calorie target in five minutes, for nothing

The most commonly paywalled item in this category is a number you can calculate yourself. Mifflin-St Jeor gives you a resting metabolic rate:

Men: (10 x weight in kg) + (6.25 x height in cm) - (5 x age) + 5

Women: (10 x weight in kg) + (6.25 x height in cm) - (5 x age) - 161

Then multiply by an activity factor. This is the step where most apps quietly lose accuracy, because they offer three choices when the formula defines five:

• 1.2 sedentary: desk work, little deliberate movement
• 1.375 lightly active: light exercise one to three days a week
• 1.55 moderately active: moderate exercise three to five days a week
• 1.725 very active: hard exercise six or seven days a week
• 1.9 extra active: physical job, or twice-daily training

The gap between 1.2 and 1.9 on a 1,500 kcal resting rate is about 1,050 calories a day. Compressing that range into three buttons means the people at either end get a target that is wrong by hundreds of calories, and the error repeats every day for as long as they use the app. When I built the target engine in Numi I kept all five multipliers for exactly this reason: the middle of the range is well served by a three-option picker, and the ends are not, and the ends are where people give up because "the numbers do not work for me".

For a loss goal, subtract from that maintenance figure rather than picking a number off a chart. Then leave it alone for a fortnight.

And track weight as a trend, not a reading

You can do this in any spreadsheet. Weigh yourself daily, average the last seven days, and only look at the average. Daily weight is dominated by water, sodium, glycogen, digestive contents and time of day. A seven-day average moves for reasons that are actually about body composition.

This is also the reason a target should not be recalculated every time the scale moves. An app that adjusts your calories daily feels attentive and is mostly reacting to noise, and the visible consequence is that people stop believing the number. Change a target on a persistent shift, meaning a trend that has held for a week or more, not on a single reading.

Where a subscription is actually earning its money

I would rather be straight about this than pretend everything should be free. A recurring charge is defensible when it is paying for something with a recurring cost, or for work that has to be redone continuously:

• Per-photo or per-message compute. Real money, spent each time you use it.
• Database maintenance. Food data rots. Products reformulate, restaurants change menus, entries get corrupted by crowd editing. Somebody has to keep at it.
• Insight work that is specific to you. Not "great choice!", which is the cheapest possible output and destroys trust precisely because it obviously is not about you. Specific means referencing what you have already eaten today, what is left, or something you have repeated four days running. That requires either per-use compute or a lot of careful deterministic engineering.

What a subscription is not earning: access to your own history, the ability to change your activity level, or a formula published in 1990.

A five-question test before you install anything

1. Can I export my data? If no, stop here.
2. Is there a history limit in the free tier? Any limit on viewing your own past entries is a retroactive paywall waiting to happen.
3. How many activity levels does onboarding offer? Three suggests the target is approximate by design. Five suggests somebody looked at the formula.
4. Does it show a weight trend or a daily number? A daily number will make you feel worse and tell you less.
5. Which features involve a photo, a chat box, or a microphone? Those sit below the marginal cost line. Assume they are capped, and find out where the cap is before you build a habit around them.

If a free app passes all five, use it and stop looking. If it fails the first two, no feature list makes up for it.

The uncomfortable trade-off nobody puts in the comparison table

Free, no subscription, and no manual food search cannot all be true at once right now. Manual search is cheap to run, which is why the genuinely free-forever calorie counters are database-and-search apps. Photo recognition removes the searching, which is the part people actually hate, and it costs money every time it runs.

So the real question is not "is it free" but "which cost am I willing to pay": a few seconds of typing per meal, or a few pounds a month. Both are honest answers. What is not honest is an app that promises the second and quietly delivers a degraded version of it, or that lets you build a month of history and then charges you to look at it.

Pick based on which cost you will still be willing to pay in six weeks, because that is the only horizon at which any of this data becomes useful.