Food Journal Apps: Words, Photos or Numbers?

I judge a food journal by week six: is it still getting opened? In a six-month trial of 128 adults, the smartphone app group recorded food on a mean of 92 days. The paper diary group managed 29 (Carter et al., 2013). However you log, the app ends up turning your words or your picture into foods and grams, then multiplying by a database.

What are the three ways a food journal app records a meal?

Words, photos and numbers. With words you type or say what you ate, like “oatmeal and a banana”, and the app matches each food to a database entry. With a photo the app names the foods and suggests grams. With numbers you scan a barcode, search a food or type the calories yourself.

Method	What you do	Where the calories come from	Where the error comes from
Words	Type or say the meal	A database entry for each food	Grams you leave out, a wrong match
Photo	Snap the plate	A database entry for each food the app names	Grams the app guesses, oil you can’t see
Numbers	Scan, search or type calories	The label, the database or you	The portion you pick

All three end in the same multiplication. A database supplies one half and the grams you log supply the other. Apps differ most on that second half, so it’s the first thing I check in any app.

How much does the way you log change the number?

A lot, and almost all of it comes from the grams. The formula is calories = kcal per 100 g × grams ÷ 100, so I ran one breakfast three ways: cooked oatmeal, a banana and peanut butter. The database numbers stay put. Oatmeal cooked with water has 71 kcal per 100 g. A peeled banana has 88 kcal per 100 g, and smooth peanut butter has 598 kcal per 100 g. The records list a cup of cooked oatmeal as 234 g, a banana as 115 g and 2 tablespoons of peanut butter as 32 g.

How it was logged	Oatmeal	Banana	Peanut butter	Total
Words, default portions	234 g = 166 kcal	115 g = 101 kcal	32 g = 191 kcal	458 kcal
Photo, grams guessed low	175 g = 124 kcal	115 g = 101 kcal	16 g = 96 kcal	321 kcal
Weighed, a deep 1½-cup bowl	351 g = 249 kcal	115 g = 101 kcal	32 g = 191 kcal	541 kcal

The photo row is an illustration of a low guess, not a measurement of any app. Even so, grams alone open a 220 kcal gap on one breakfast. A journal that hides the grams hides exactly the number that changes the total.

Weigh the two or three foods you eat most, once, and save them as your own foods, because in the table above the oatmeal and the peanut butter did the damage and both are easy to eyeball. A scale helps most in the first week. Then your saved foods take over.

Why do dry and cooked weights give such different numbers?

Water. The USDA record for raw brown rice gives 367 kcal per 100 g, and the record for cooked brown rice gives 123 kcal. A cup of the raw grain is 185 g and a cup of the cooked is 202 g, so the two cups hold 679 kcal and 248 kcal. Same rice, same cup, 431 kcal apart.

Pasta behaves the same way. The record for dry pasta says 371 kcal per 100 g, and cooked pasta says 158 kcal. A common slip is weighing 75 g of dry pasta and picking the cooked entry. The diary shows 75 × 1.58, which is 119 kcal, when the dry entry gives 75 × 3.71, which is 278 kcal. About 160 kcal vanish. That is 57% of the real total, and nothing on screen looks off.

Log the state you weighed. I weigh dry, because I know what went into the pan and the portion is the same every day. The paper template below has a column for the state of the food for this reason.

What do I write for a restaurant meal?

The dish, the place and your best guess at the grams, marked as a guess. Menu numbers are a fair start. A 2011 JAMA study measured 269 items from 42 US restaurants with bomb calorimetry and found that stated and measured energy did not differ significantly on average. But about 19% of the items held at least 100 kcal more than stated, and the understated ones tended to be lower-energy foods.

So a salad listed at a modest figure may hold more than the menu says. Write “menu” next to that line and move on. Later, the lines marked “menu” and the lines marked “weighed” show you which days were firm and which were sketches.

How accurate is a photo log?

In one small pilot it ran low, and so did the standard research method. A New Zealand team tested an iOS photo app (Lee et al., 2025) against energy expenditure estimated from indirect calorimetry and a wrist accelerometer. Over seven days the app’s intake estimate sat a mean of 1,814 kJ (about 434 kcal) a day below that figure. A 24-hour recall sat 1,715 kJ (about 410 kcal) below. The authors judged the app comparable to the recall.

Only 26 people had app and expenditure data, and 18 had the recall. For one person the daily gap could land anywhere from 7,629 kJ low to 4,001 kJ high. That’s wide. I read one day as an estimate and a week as a better one.

A photo is a quick first draft, and the grams are yours to confirm. An app that makes you confirm them before saving earns a point from me. The same limit shows up when a chat reads your diary, as the post on what an AI nutrition coach app can and cannot do explains.

What goes in a paper food diary template?

Five columns: time, food and its state, grams, kcal per 100 g, and kcal. The state matters because cooked and raw weights differ, and the grams have to match the state in the record. Here is part of a day, filled in from USDA entries.

Time	Food and state	Grams	kcal per 100 g	kcal
7:30	Oatmeal, cooked with water	234	71	166
7:30	Banana, peeled, raw	115	88	101
13:00	Chicken breast, roasted	140	165	231
13:00	Brown rice, cooked	202	123	248
16:00	Almonds	28.35	579	164
	So far			910

The almonds are 1 oz, which the record gives as 28.35 g. Each kcal figure is kcal per 100 g × grams ÷ 100, rounded. Add it up at night. A running total during the day invites haggling over dinner.

Keep it neutral. Write what you ate and how much, with no verdict on the food and no column for guilt. If writing everything down has started to feel hard to stop, put the notebook away for a week and talk to your doctor.

How long should I keep a food journal?

Longer than one day, and the reason is day-to-day variation. A 1987 USDA study followed 29 adults for a year and worked out how many days of records land within 10% of a person’s usual intake. For food energy, a group’s average needed 3 days. An individual needed 31 days on average, which is a month of records. One day tells you what Tuesday was. It doesn’t show your usual week.

Weighing the foods you eat most is the cheapest place to start.

My routine is a full week, then I decide what to keep. Most of my breakfasts repeat, so those become saved meals and I only weigh what changes. The first week is for learning your own pattern, and the days that look alike show up fast.

How do I choose a food journal app I will still use in week six?

Time is the cost you pay every day. In an online 24-week program, 142 adults spent about 23 minutes a day logging in month 1 and about 15 in month 6 (Harvey et al., 2019). Even the lower figure is a real slice of a day. I choose the app that passes these seven checks and don’t care how pretty its charts are. The ten-test version is in how to choose a calorie counter app.

1. It records by words, photo and numbers, so a rushed meal still gets in.
2. It shows the grams and the database source behind every number.
3. It lets you correct a gram figure after the fact.
4. It repeats a meal or a whole day in one tap, and copies only the foods you pick.
5. It works without a connection.
6. It lets you read and export your diary after you stop paying.
7. It treats a skipped day as nothing, with no streak to lose.

To test those seven in ten minutes, use the free tier or the trial, a scale and yesterday’s oatmeal from the table above.

Check	What to do	It passes if
Words, photo, numbers	Log breakfast by typing, lunch by photo and a snack by barcode	all three save an entry
Grams and source	Open any entry	the grams, the kcal per 100 g and the source name show
Correcting grams	Change the oatmeal from 234 g to 351 g after saving	the total moves from 458 to 541 kcal without starting over
One-tap repeat	Copy yesterday’s breakfast	it takes one action
Offline	Switch on airplane mode and log a meal	it saves and syncs later
Export	Find the export button before you enter data	the file opens in a spreadsheet with foods, grams and dates
Skipped day	Skip a day, then reopen the app	no streak or scolding

If ten minutes is too tight, do the grams, the correction and the offline rows first. They decide whether the numbers can be trusted when you’re tired.

A review of 22 studies found a consistent link between self-monitoring and weight loss (Burke et al., 2011). The authors rated the evidence weak, because of how the studies were built. I won’t promise that a journal makes you lose anything. What it does give you is a clear view of what you ate.

If you’re thinking of leaving a Points-based app for a calorie counter and want to know what the move costs you, the guide to moving from Points to calories covers the monthly fee, how cancelling works and how to get your data. If you’re leaving a large-database tracker, what to check before you switch starts with export and what your plan includes.

In Calsa you can type or say a meal, take a photo and confirm the grams, scan a barcode, photograph a label or type calories, and each entry shows its source and the “per 100 g × grams” math. Copying a day, logging offline and exporting everything all work, and there are no streaks. It’s a paid subscription with no watch app and no weekly meal plans. The paper template above costs nothing.