A/B testing,
built with AI.
Describe a test in plain English. Splitty builds it, runs it on your store, and tells you what won.
The problem
Testing is slow, technical and confusing.
Most store owners never run a test. Setting one up needs a developer, and reading the results needs a statistician.
The idea
Just say what you want to try.
Type "make the add-to-cart button bigger." Splitty edits the live page, sets up the test, and explains the result in plain English.
How it works
Three steps.
1. Describe it
Tell Splitty what to change. It edits the real page while you watch.
2. Launch it
Pick who sees it and split the traffic. Checks catch problems before you go live.
3. Read it
See the winner, the revenue impact and how sure you can be, no stats degree needed.
Splitty
An AI assistant that edits the page in place.
I designed and built Splitty end to end, the chat, the live editing, and the ideas it suggests for each page.
The dashboard
Everything a test needs, one screen at a time.
Results that read like a sentence
"Variant B is 96% likely to win." Then the detail, if you want it.
Revenue and funnels
See where money moved, not just clicks.
Alerts
Flags odd traffic or broken variants early.
AI analysis
A written summary of what happened and what to try next.
Ideas from Splitty
Test ideas grounded in your page.
CSS, JavaScript or redirect
Change anything on the page.
Precise targeting
Pages, devices and audiences.
Traffic split
Even splits or custom weights.
Anomaly alerts
Problems flagged while tests run.
Test Plans Pro & Agency
An AI-planned roadmap.
What I did
The parts you see and touch.
Branding, every screen of the app, the front-end build, and Splitty. The backend and stats engine were built by the engineering team.



