A Quit-Sticks-and-Vape Tracker Built via Vibe Coding as a Browser Web App

⚡ The Gist in 5 Seconds - A project manager with no programming experience built a quit-sticks-and-vape tracker in a dialogue with the Claude neural network: the model writes the code, the author sets the task and checks the result. - It is a web app on static hosting with no backend — it opens in the browser and can be added to the home screen of an iPhone, Android device, or computer. - Limitation: the app makes no diagnoses and does not replace a doctor; no personal data is collected. ### 🔍 What Was Found The case study was published on Habr on October 1, with a reading time of 8 minutes and 5.9K views. The author has worked in digital agencies for over a decade: as an account manager and then project manager — websites, CRM marketing, design, campaigns for FMCG, pharma, retail, and the public sector. When she was quitting smoking, existing apps tracked days and money saved well, but they were built around regular cigarettes: nothing could be found about sticks and vapes, and the most effective feature — body recovery progress — was about cigarettes and tar. The app has three timelines — for cigarettes, sticks, and vapes — and next to each milestone there is a letter indicating the level of evidence: A — large studies and meta-analyses; B — individual studies or the direct effect of nicotine, which is the same for all three groups; C — data extrapolated from cigarettes where the device may change the picture and no direct studies exist. Key sources: CDC and NHS guidelines and Surgeon General reports, the 2018 US National Academies review on e-cigarettes, brain imaging studies by Cosgrove (2009) and Rademacher (2016), Taylor's meta-analysis in BMJ (2014) on mental health after quitting, Hughes' (2008) work on relapses, and Nair's (2003) gum study. The first version was a native iOS app in SwiftUI: two full code iterations hit a wall not in the code but in publishing — a paid Apple Developer Program account, payment and identity verification from Russia as a separate project with its own timelines and risks, plus store moderation. For a free pet project, that was disproportionately expensive, so everything moved to the web. To make the site behave like an app on iPhone, a few meta tags are enough: fullscreen mode, the name under the icon, status bar color, and viewport-fit=cover stretches the UI around the notch. ### 💡 Why It Matters The difference between devices is not minimal: in sticks, tobacco is heated rather than burned, and vapes contain no tobacco at all — meaning some items on the cigarette timeline apply to them differently. But nicotine-related aspects — cravings, sleep, receptor function — are roughly the same for all three groups. At the same time, there are noticeably fewer studies on heated tobacco and vapes than on cigarettes, so the author did not want an app that pretends to know exactly what is happening to the lungs in week three. The evidence levels next to the milestones are an attempt to show where there is knowledge and where there is an educated guess; the medical disclaimer is presented separately. ### 🧩 Context Vibe coding is when you describe to a neural network what you need, and it writes the code. According to the author, her role is almost

🔗 Read on habr.com