Gymi: a personal gym coach on WhatsApp
A Hebrew coaching service that turns a gym membership into a plan. Every plan is built by a method a certified fitness coach designed, you ask on WhatsApp what today looks like, and the coach remembers last session's numbers. Co-built with that coach, out of a need we both had, and now in a closed pilot fed by a waitlist that runs itself.
The product is public: gymi.co.il
Trainees' conversations are private and stay private. Everything below is product, method, and go-to-market.
Where it came from
It started as our own problem. Most people paying for a gym membership do not lack access or willpower. They walk in without a plan, repeat last week, and stop progressing. A personal trainer fixes that and costs more than the membership. An app hands you a plan and leaves you alone with it.
We wanted the thing in between, for ourselves first: a plan built for this gym and this goal, and someone to ask "what do I do today?" in the place we already are. In Israel, that place is WhatsApp.
Two builders, one clear split
Gymi is built with a certified fitness coach. He built the training engine and the method every plan is built from. I built the product around it - the WhatsApp channel, the onboarding conversation, the positioning, and the go-to-market.
That split is also the product's central claim, and the wording is exact on purpose: the engine and the method were built by a certified coach, and every plan follows that method. It does not say a coach reviews each plan, because one does not.
A promise list, not a pitch
Before any marketing was written, every public claim was checked against what is actually built and sorted into what we may say and what we may not. "A video for every exercise" became "demonstration videos for every exercise". A rest timer that does exist is described as rest time after each set, nothing grander. Human support stays off the list until someone is named to answer and a response time is set.
The landing page and the ads are written from that list, and the coach does not quote prices, free trials, or medical advice.
The conversation is the product
Joining the programme is one message: the goal, how many times a week, and what the gym has. Back comes a plan built for that gym. After that, the day-to-day is a chat:
"What are we doing today?"
"Workout B, about 50 minutes. Squats first: 4 sets of 8. Last time you did 57.5 kg - today we try 60."
The coach remembers the last session, so progression is a conversation rather than a spreadsheet. The workout itself opens from a link, with demonstration videos for every exercise and rest time after each set. Any question goes back to the chat.
Built for beginners, on purpose
The engine is built for someone starting out. Level is part of the profile, and advanced exercises stay locked until the ability is shown. Progress is earned, not assumed - which is exactly what a good coach does in a first month.
Go-to-market as part of the product
- One page serves every ad. The ad picks the headline, and its code travels into the WhatsApp message, so every sign-up arrives already attributed to the ad that brought it.
- Joining is one message. You are on the list first, then answer three tap questions that help choose the pilot group. Typing is never required.
- Leaving works at any point, with one word, and after that nothing more is sent.
- The pilot is chosen from real answers, not from a form nobody finishes.
The parts that are not the happy path
- Not a medical service, and the product says so where it counts.
- Deleting the account deletes the trainee - the number, the conversations, and the coach's memory. The privacy policy names the two records kept on purpose, and why.
- Plan creation has a per-account quota that deleting and signing up again does not reset.
- It runs on Meta's official WhatsApp platform, around the clock, not on someone's laptop.
Outcome
A product in closed pilot: plans built to a certified coach's method, a coach on WhatsApp that remembers, and a public site and waitlist that sort and attribute sign-ups on their own.
What it demonstrates: turning a personal need into a product with a partner who owns the domain, keeping every public claim provable, and building the go-to-market as part of the product rather than after it.