A pet-health app I designed and built alone, from the first sketch to production: health records, reminders, an AI assistant that knows your pet, emergency guides and a lost-pet QR tag, in 14 languages. Live at bpetit.com since the end of August 2026, with an Android app in testing on Google Play and a connector that lets AI assistants read and update a pet's record.
Why it exists
Owning a pet means keeping a dozen things straight: vaccines and their boosters, the vet's last notes, the weight curve, the pills and their doses, the papers. They usually live in a drawer, a chat and someone's memory, and they are never there at two in the morning when the dog has eaten something it shouldn't.
Pet apps tend to do one slice of it: a social feed, a clinic's records, a reminder list. BPetit puts the whole record in one place and adds the two things you need when it goes wrong: someone to ask, and clear steps to follow.
What it does, in four videos
Four short explainers, made from real captures of the live app with a demo pet called Luna.
Beyond the record
Pets can be shared with a co-owner, or with a sitter for a set number of days. Reminders reach the phone as notifications and as an Android home-screen widget. A premium plan is in place. And BPetit is also a connector for AI assistants: through its MCP server, with its own secure sign-in, an assistant like Claude can read a pet's record, log a vet visit or mark today's dose, with the same permissions the owner gave in the app.
Designed and built alone
I did every part of it: research, product, brand, interface, code, data, payments, store listings. The first version, in 2024, ran on MongoDB and taught me what not to do. I rebuilt it from scratch on Supabase, shipped version 1.0 in April 2026 and opened it to the public at the end of August.
Today it runs on Next.js 16 and React 19, with Supabase for data and sign-in, Paddle for payments, Resend for email and Sentry for errors, hosted on Vercel. The AI runs through an open, low-cost model. The interface works in 14 languages, Arabic, Hebrew, Persian and Urdu included, written right to left. The Android app is a native shell with push notifications, now in closed testing on Google Play.
What it taught me
AI-assisted coding changes who gets to build. I'm a designer, not a trained software engineer, and BPetit is a production app with payments, security and a store listing, built by one person.
Design sense carries into code. Knowing what the user should feel at each step makes it obvious what the code has to do.
Scope is the hardest discipline. There is always another feature. The work is deciding what the product is for and finishing that first.
See it at bpetit.com.