How to count calories without weighing food

The food scale is the gold standard of calorie tracking — and the number one reason people quit. Weighing every ingredient turns each meal into a chemistry experiment, and almost nobody sustains it past the first burst of motivation.

The good news: you don't need gram-level precision to lose, maintain, or gain weight. You need a consistent estimate of your intake, sustained over weeks. Here are four ways to get it without ever touching a scale, ordered from lowest-tech to lowest-effort.

1. The hand portion method

Your hand is a portable, personal measuring tool:

• Palm = one portion of protein (~25–30g protein, roughly 120–160 kcal for lean meats)
• Cupped hand = one portion of carbs (~25–35g carbs)
• Thumb = one portion of fats (~10–12g fat)
• Fist = one portion of vegetables

Count portions instead of grams. A meal of two palms of chicken, one cupped hand of rice, and a thumb of oil is a fully “tracked” meal — no numbers needed. It's crude per-meal, but because your hands don't change size, your errors are consistent — and consistent errors are correctable: if the scale isn't moving, remove one daily portion and reassess in two weeks.

Best for: people who want zero apps and zero logging. Weakness: no running total, no macro detail, easy to “forget” snacks.

2. Photo logging with AI

Take a picture of your plate; an AI identifies the foods, estimates portions, and logs calories and macros. What used to take a minute of database searching takes five seconds.

Is a photo estimate as accurate as a scale? No — expect meaningful error on individual items, mostly from portion-size guessing and invisible cooking fats (we've written an honest breakdown of AI calorie counter accuracy). But it wins where it counts: because logging is nearly effortless, you log everything, every day. A complete week of good estimates beats a half-logged week of perfect ones — and incomplete logging is precisely how manual tracking fails.

Best for: anyone who quit MyFitnessPal-style logging out of tedium. Weakness: underestimates oils and sauces — add a quick note for those.

3. Voice logging

Say what you ate: “two eggs, a slice of buttered toast, and a coffee with milk.” Speech-to-nutrition logging parses the sentence, matches each item, and applies standard portions.

Voice shines exactly where photos fail — invisible and non-visual items. The latte you drank in the car. The handful of almonds at your desk. The cooking oil the camera can't see. It also removes the last excuse for not logging: you can do it while walking out of the restaurant.

Best for: drinks, snacks, restaurant meals, and people who hate typing. Weakness: “a bowl of pasta” still requires the AI to assume a portion — say “a large bowl” and it gets closer.

4. Calibrate once, then repeat

Most people eat the same 10–15 meals on rotation. That's not boring — it's a tracking superpower. Carefully estimate each recurring meal once (this is the one place a food scale earns its keep for a single evening), save it, and from then on log it in one tap.

Combined with method 2 or 3 for novel meals, this gets you scale-level accuracy on 80% of your diet with near-zero daily effort.

How accurate is “accurate enough”?

Here's the reframe that saves most trackers: your goal is a reliable weekly trend, not a perfect daily number. Research on dietary self-reporting shows even diligent manual loggers misestimate intake substantially — so chasing perfection was never on the table. What works is a method whose errors are consistent, applied every single day, checked against your weight trend every few weeks. All four methods above pass that bar. A scale used for two weeks and abandoned does not.