I’ve been weighing myself every morning for most of this year. The number goes up. The number goes down. I want it to go down more often than up. That’s basically the whole story.
To do that consistently, I need to know what I’m eating. And every calorie tracker I tried felt like it was designed for someone with more patience than me – apps with onboarding flows, dashboards, paywalls, ten taps to log a banana.
I just wanted to type “two eggs and toast” and be done. Or take a pic. Or dictate!
So I built it.
What it is
BiteNote is a calorie and macro tracker that lives inside Telegram. You message @bitenotebot the way you’d message a friend:
two scrambled eggs, slice of rye bread, black coffee
It figures out the food, estimates calories and macros, logs it, and shows you where you are for the day. No app to install. No account to create. No dashboard to visit. The conversation is the interface.
Behind the scene it either checks the database or tries to get estimates from gemini. It might be wrong! But also when I’m trying to eyeball calories I might be wrong too, so it’s guess is as good as mine (probably better though).
The problem I was actually trying to solve
The problem isn’t that calorie trackers don’t exist. There are dozens. The problem is the friction between deciding to log a meal and the meal being logged.
If logging takes 30 seconds, I do it. If it takes 90 seconds and three taps, I skip it, then skip the next one, then stop opening the app for a week. And then suddenly I find myself with gaps in my logging and then I drop the app and weight goes up because I’m like this.
Telegram solves this for me because:
- It’s already on my home screen.
- I’m already in it ten times a day.
- Typing a sentence is faster than navigating any UI.
- An LLM on the other end can parse “kanapka z serem i pomidorem (cheese and tomato sandwich but in Polish)” without me picking from a dropdown.
Why
Two reasons.
One – I wanted it for myself. I’m 94kg, I want to be lighter, and I know the only reliable way to do that is to keep food intake honest with myself. Nothing on the market made that easy enough that I’d stick with it for more than two weeks. And I have tried MacroFactor for the last couple months!
Two – I wanted to see if a chat-native product could replace an app-native one. I’m a PM by day. I spend a lot of time thinking about what AI changes about the shape of products. BiteNote is my own answer: for a category like food logging, the app is (?) the wrong unit. The conversation is the right unit.
It started as a weekend project. Then someone paid for it. Then a few more people did. So now it’s a thing.
Also, all (or most) of us have some kind of chat-interfaced AI’s that we use everyday, so using the same interface felt right.
What’s there now
- Natural-language meal logging,
- Daily totals, macros, weight tracking,
- Photo logging — send a picture, get an estimate,
- Restaurant menu lookups,
- A free tier that’s actually usable, and a Pro tier for people who want history and integrations (there’s Withings integration and API access!).
It’s at bitenote.eu if you want to poke around. The bot is @bitenotebot on Telegram.
What’s next
I’m building this in the open (or this blogpost is actually a start of that), slowly, on the side. The roadmap is whatever annoys me most when I’m using it. Right now that’s better weekly summaries and tighter integration with the scale data I already pull from Withings.
If you try it and something breaks or feels stupid – let me know!