World dominationstarts with lunch.

Big plans. First, lunch

I came to this from business: more than fifteen years in sales, negotiations, business development and managing teams. I had always wanted to build something of my own. Then AI offered a way to make software myself, even though I barely knew how to program. How could I resist?

I did not have to look far for a problem. I had a background in powerlifting, bodybuilding and preparing for competitions. I had counted calories plenty of times. With existing diaries, I either disliked the design or spent too long figuring out how to record an ordinary meal. Sometimes a spreadsheet seemed easier.

So I made a Telegram bot. Enter a food and a portion; get calories, protein, fat and carbs; keep the entry in a diary. A modest beginning for someone with plans for world domination. But at least I could log lunch.

Now make it all work together

One entry soon stopped being enough. I wanted to see the whole day, adjust a portion and save favourite foods. The bot gained screens: a Mini App inside Telegram. Meal plans, reminders, Russian and English followed. Each addition seemed perfectly reasonable. On its own.

Putting them together was more interesting. Adjust a portion, and you need to understand what happens to the daily total. Change one screen, and the next needs checking too. Working with GPT and Claude regularly turned into a programming lesson for me. “Write a prompt, get an app” began to sound particularly optimistic.

I learned to ask questions before agreeing with an answer. What happens to an existing entry? What if someone closes the screen? Can they export their data or delete it? The diary was accumulating a history, and that history needed care.

The project gets a name

Gradually, I realised I was building more than a calorie-counting bot. The name Sanvimo brought together health and energy. I like the idea of having enough energy for sport, work and projects of your own. I would like to keep my favourite food in that life too.

Then came Sanvi, a silver robot with a green eye. Curious, a little funny, on your side. For now, Sanvi welcomes people on the website and reacts to actions in the demo. An independent AI assistant is still some way off, but the personality is already there.

The mountains and stars stayed here on the roadmap. Sometimes I like to look up from the button I am working on and think about where all this might lead.

The saucepan has a say

Home cooking quickly explained why a food database alone was not enough. Ingredients go into a pan, you cook, then weigh the finished dish. Some water evaporates; some gets added. Then you serve yourself only part of it. That is the part you need to log.

Sanvimo gained a personal recipe book, container weight and portion calculations based on the finished dish. Save the ingredients and use them again. Take a recipe from the shared collection and cook it yourself. Authors, follows and conversations began to grow around the food.

I was curious about something else too: what does the food on my plate cost? Pro gained calculations using your own prices. A portion’s cost and a supermarket receipt are different amounts, though. You buy a whole bottle of oil and use a little. Shopping by the package is a question for another chapter.

The free recipe that wanted payment

I had decided that recipes should be available without paying. Let someone cook a dish, log a portion and find out whether the app suits them. Then my phone showed: “Creating and editing requires Pro.”

Excellent. I had decided recipes were free. The app had other ideas.

We investigated and found a dependency on the part of the app that checked payment access. If it returned an error, free recipes could fail to load and an old restriction could appear. We could not establish exactly what happened on the phone that time. But we reproduced and fixed this failure.

Earlier, another case had let someone start a recipe only to refuse when they tried to save it. Spend your time first; discover it was wasted afterwards. My ideas about usability had met my own app. It was a useful encounter.

After that, we began walking through the whole journey with and without Pro. Free lets you create up to ten recipes, edit them and log portions. Limits appear upfront. I want a first experience of Sanvimo to start with a home-cooked meal, without that kind of surprise.

An icon of my own

After Telegram came browser access, then Android. It feels good to see your own icon beside the apps you use every day. Once, it was just an idea you discussed with AI. Now you tap it, and it opens.

I really did not want someone to find an empty diary behind their new icon. So we worked on a shared account: sign into it another way and continue your entries. Changing screens was supposed to make life easier, without making anyone start over.

Android Beta has been available from the website and the Sanvimo channel since 17 October. On 18 October, I installed build 16 over the previous version and signed in. My history was still there. A wonderfully uneventful result. Exactly what you want from an update.

Later builds keep arriving separately. Each still needs its own phone checks; one successful installation cannot speak for all the next ones. The app now has a notification bell. Stores, push in the closed app and automatic updates remain ahead.

Real money. My own purchase

My business background was hardly going to let me forget about money. I want to keep developing Sanvimo, and the project needs to support itself. People should still get something useful for free: log food, work out a recipe and come back the next day.

On 16 October, I bought my own Pro on a separate account. The special test terms were ten rubles for 360 days. Regular Pro has a different price. The money was real; the buyer was me. Two short messages appeared: “Payment confirmed” and “Access granted.”

A lot of work sat behind those messages. And new questions followed them: where is the purchase, where does the code go, what needs refreshing? In a business conversation, I can notice someone hesitating and explain. Here, I am not sitting beside them.

So we returned to checkout. A discount lowers the price; a gift adds Pro days. What a code will do should be clear before anyone presses the button. Purchases use Stars in Telegram and YooKassa in the browser. If buying is unavailable, that needs to be clear straight away. Receipt and email delivery have their own checks. A working payment turned out to be the start of another chapter.

The everyday test

The longer I work on Sanvimo, the more often I return to small things. A lovely website background jumps when you scroll. Recipes are explained, yet it is still unclear which weight to enter. A link meant to open purchase history in the browser sends you into Android.

One continuous landscape solved the background problem. An example on the site made portions tangible: take 200 grams from an 800-gram finished dish, and you get a quarter of its calories. Change the weights and watch the calculation. Free terms appear upfront, with sign-in, Telegram and Android in the header.

The same attention helped inside the app. Show a photo before saving it. Do not let a recipe-name error take the whole draft with it. Clear unsent analytics when someone signs out. Some of these improvements have shipped; others are still being checked.

Things like that take up quite a bit of my time now. From the outside, a button got fixed. From my side, I learned a little more about how someone would use what I had made.

What weight did I use last time?

After food, my thoughts naturally returned to the gym. Between sets, I want to record weight and reps quickly, then see where I left off next time. The personal diary is now in the app: a day, your own exercises and sets. Trying it on a phone immediately gave us more work, especially entering numbers and returning to the right screen.

A whole training week has now grown around those entries. Save a good session as a template, build a programme and put it in the calendar. Plans change; move the workout. Programmes, templates and scheduling are now available. An ordinary week fits in Free; longer and less usual cycles are part of Pro.

History lets you revisit earlier sets. Pro adds charts and period comparisons, so you can follow load and consistency. It also estimates calories burned. Estimates is the key word. There are no watch readings behind that calculation yet, and it will not automatically add food to your daily target.

Achievements have arrived too: a first session, coming back after a break, a personal result. I have no interest in extra weight on a bar just to earn a badge. A reason to notice that you are keeping going sounds much better. The new Android build still needs its own phone checks; updating the server cannot do those for us.

Sometimes you need a person

A diary remembers a lot, but sometimes you want to talk it through with someone who can help. At first, I pictured a coach workspace. Now I am looking more broadly: coaches, nutrition specialists, physiotherapists and doctors. Each has a different job; the person’s history is still one history.

That idea became a working workspace. Weekly summaries were followed by a shared plan. One sign-in, with separate screens for your own diary and your work as a coach. Either person can propose a goal and steps, discuss changes and keep a history of what they agreed. A new plan takes effect when the other person accepts it. The workspace has moved into open Beta, with friendships and coaching requests. Clients still choose the data, permissions and access period.

I wanted meals and workouts to be just as clear. A coach proposes food or sessions for particular days; the client looks through them and decides. Accepting puts them in the ordinary diary and calendar. From there, the distinction matters: a scheduled workout has not been completed, and a planned lunch has not been eaten. You can leave a short report, ask a question about the assignment and get an answer. The numbers begin to have some context.

Then came another question: how does a coach keep track of the person behind all those entries? The queue now shows who needs an answer and which reports still need a review. Set a date to return to someone; a new question will stay visible in the meantime. Opening a report and reviewing it are separate actions. The Beta now includes a services catalogue with test credits. Real service payments are not available, and medical consultations remain planned.

A good programme can travel further

Suppose you have made a programme you like. A friend asks for it. You should be able to send one link: they save their own copy, while your personal results stay with you. That is now how it works: sharing is free, and every copy has its own life from there.

My next thought is fairly predictable: if a specialist has put their experience into that programme, why not let them offer it for sale? A buyer should see what they are paying for, and a creator what they have earned. The Beta now lets people try the services catalogue and balance with test credits. It takes no real money and pays none out. Real funding, refunds and payouts still need to be worked out. It is a substantial undertaking, but it could make useful experience easier to access and give its author a way to be paid.

Plenty of numbers, not much clarity

There is another area I want to explore: lab results. A report arrives, full of abbreviations, numbers and marks. You want to keep it, understand the names of the tests and take useful questions to your doctor.

We are exploring how to help make sense of it: keep the original, carefully extract values for your confirmation, explain the terms and review changes using the laboratory’s own units and reference ranges. You would be able to share a selected document with a specialist separately. Diagnosis and treatment remain a doctor’s work; our first task is to handle the data carefully and make the report easier to understand.

Watches, grocery shopping from a meal plan and sharing experience are still on the map too. The diary is gradually becoming a place where these parts of life could meet. Below, you can see what works today, what we are building and what we are still exploring.

I started out curious, wanting to make something of my own. Now it has a name, a diary, recipes and an icon on my phone. And I have a much better idea of what I got myself into. The plans are still big. We will just start world domination with lunch.