Photo recognition
Snap any meal. GPT-4 Vision identifies the food and portion, returns a full nutrition breakdown.
Send a photo of your meal to Leam — our AI calorie counter in Telegram returns calories, protein, fats and carbs in 3 seconds. Calorie counting with no forms, no databases, no manual entry.
Free to start. No credit card. 1,500+ people tracking with Leam.
Leam is a Telegram-native calorie tracker built for humans, not spreadsheets.
Snap any meal. GPT-4 Vision identifies the food and portion, returns a full nutrition breakdown.
Say "pasta with chicken for lunch" — Whisper transcribes, AI logs. Hands-free while cooking.
No forms. Talk to the bot in plain English or Russian. It understands context, portions, units.
Daily weight logging, Apple Watch-style rings for calories in vs out, streaks and progress charts.
Forgot to log yesterday's pizza? Just say "yesterday I had pizza" — Leam handles the date.
No install. No account. Open Telegram, tap the bot, start tracking in 10 seconds.
Three steps. Under 30 seconds to your first meal logged.
Tap "Open in Telegram" — our bot starts the conversation and asks a few questions to set your calorie target.
Photograph your plate, record a voice note, or type "200g rice + grilled chicken". AI does the rest.
Get instant calories, macros and an Apple Watch-style ring. Weekly trends and weight charts come automatically.
Calorie counting fails for one reason — friction. Leam removes it.
Most diet apps die on day five. The reason is boring and consistent: calorie counting is a 40-second workflow in the best apps, and after a few days that workflow feels like a part-time job. Photo recognition and a chat interface change the math.
Leam is an AI calorie counter built as a Telegram bot. You send a photo of your meal, the AI identifies the food and portion, and you get calories plus a full macro breakdown — protein, fats, carbs — in about three seconds. No food database to search. No serving-size confirmation modal. No ritual of opening an app before every meal. The friction that makes traditional calorie counters an exhausting daily chore is simply gone.
Under the hood, Leam uses GPT-4 Vision for food recognition and backs its calorie math with real nutrition science. Your daily target is computed from Mifflin-St Jeor BMR and a TDEE activity multiplier, then adjusted weekly against your actual weight trend — we don't pretend a fixed calorie number works forever. Read the full methodology in our guide to BMR, TDEE, and calorie deficits: see the blog posts linked below for details on the formulas and the research they rest on.
Telegram is where you already are. That means Leam works on iPhone, Android, Mac, Windows, the web — anywhere Telegram runs. Zero install. Zero signup. Russian and English are native, with voice input (Whisper) and photo as first-class ways to log. Retroactive entries handle natural phrasing like "yesterday I had pasta" so you're never locked into real-time tracking. The whole product is designed around one assumption: you'll only log meals consistently if logging is faster than thinking about logging.
In-depth takes on AI tracking, weight-loss science, and the apps we're trying to replace.
A glass of wine isn't the end of the world. But it does something specific to your metabolism — and the knock-on effects matter more than the calories themselves.
Drop the post-workout shaker in a panic, or you'll lose gains — right? Actually no. The 30-minute anabolic window is one of the most over-sold claims in fitness.
Photo-based AI calorie counters promise zero effort. We measured the accuracy gap against dietitian baselines on 200 real meals — here's what the numbers actually say.