01
Start from real friction.
I don't begin with market maps. I begin with something that bothered me enough to make a tool for it.
I'm Esther — a non-coder builder from China, learning product, code, and distribution by shipping real things.
I used to think building software was something only engineers did. Then AI changed the distance between having a problem and making a tool for it.
So I started building. Mostly from things that personally annoyed me — resumes I couldn't tailor, photos I couldn't pose for, astrology reports I couldn't decode, AI workflows that kept breaking.
I'm not trying to build a clean startup arc. I'm just building from things that actually bothered me.
Some of these turned into products. Some are still rough demos. A few will probably never go anywhere — and that's fine, because each one taught me how to take something messy and make it slightly more usable.
This site is where I keep the receipts: the products that worked, the demos that are still rough, the ideas I'm still testing, and the notes I'm collecting along the way.
How I build
02
Make the first version visible.
A working demo teaches me more than a perfect idea.
03
Keep the human feeling.
The best AI tools shouldn't feel like machines judging people. They should feel like someone helping you move forward.
04
Ship before it feels ready.
That's the whole point. Confusion turns into clarity faster when something is already in front of you.