How to log a Korean table dish by dish, and what this app will not do

Pick what you ate and it shows the calories and nutrients recorded for it in the Korean Food Nutrition Database. There is no account and no login, and what you log never leaves this browser.

The data comes in two kinds. Dishes are cooked food such as kimchi stew or bibimbap, and this set also holds menu figures that restaurant chains filed themselves. Packaged products are things sold in a bag or a bottle, so they carry both a product name and a manufacturer.

Four steps. For a full table, repeat them once per item.

• Choose the meal (breakfast, lunch, dinner or snack).
• Search by name. Korean initial consonants work too (ㄱㅊㅉㄱ finds 김치찌개).
• Set how much you ate: a number of servings, a number of packages, or grams and millilitres directly.
• Press Add and it joins the total. Added the wrong thing? Press Remove.

Logging rice, soup and each side dish separately is the whole method. There is no row in the database for a whole table, and pretending otherwise would be a number we invented.

The database records content per 100 g or per 100 ml, so the amount has to be in grams or millilitres. There are three ways to get there.

• By serving size - when a reference serving is recorded, we multiply it by the number of servings.
• By whole package - for when you finished the bag: package contents times the number of packages.
• Enter the amount - weigh it or estimate it in grams. This is the only figure a person sets.

Some entries record the reference serving as several forms at once, such as "레토르트 200g 기타 100g" (retort pouch 200 g, other forms 100 g). For those we do not choose for you; the option simply is not offered.

We never convert millilitres to grams or the other way round. Without knowing the density, converting would be inventing.

Different products share a name all the time, so product results always show the manufacturer alongside. Match it against the company printed on the back of the pack.

These are the figures the manufacturer declared and filed with the ministry, which is exactly why they beat anything guessed from a photo. If a product is not there, we say so - we never fill the gap with a similar product's numbers.

Most people eat the same handful of things. After a couple of days of logging, it is mostly tapping.

• Turn on on-device history and your recent and frequent foods collect, ready to add again with one tap.
• Choose the meal first, then add the items in a row; it saves a lot of tapping.
• Home cooking varies every time. Pick the entry whose preparation is closest, and if you need an exact figure, weigh it and enter grams.

Some things we cannot do, others we could but choose not to. Both are listed.

• Recognising food from a photo - no. A table set with several bowls is exactly where that approach fails hardest.
• Estimating calories - no. We state the figures in the database, and where there are none we say so.
• Dieting, weight loss, blood sugar or disease management - no. This is a tool for recording nutrition figures; health questions belong with a doctor or a dietitian.
• Targets and warnings for falling short - we do not build them. How much a person should eat differs by person, and it is not ours to decide.
• User-submitted values - we do not accept them. Values nobody can verify would make the whole dataset untrustworthy.
• Accounts and server storage - there are none. Your log lives in your own browser, and only if you turn it on.