PageDub Case Study

  1. Home
  2. PageDub Case Study
Build webpages with AI — PageDub
Make a website from a chat — PageDub
Let AI build it — PageDub
Publish and share single clean links — PageDub

Project Snapshot

What is PageDub?

PageDub is a mobile app that lets anyone build a complete webpage using AI, just by typing a description on their phone.

What PageDub Can Do

Type a description like "Create a landing page for my coffee shop"
AI instantly generates a full, styled HTML webpage — no code needed
Long-press any section to select it, then describe the change
AI edits only that component, leaving the rest untouched
Select multiple sections and update them together in one message
Publish to a live public URL in one tap — works in any browser
Save drafts, manage multiple pages, and update or delete anytime
Sign in with Google, Apple, or email — pages sync across devices
PageDub AI page generation

Problem

AI can generate page structure quickly, but taking that output from idea to a live, shareable webpage still normally means a laptop, a code editor, and a deployment step. PageDub was built to collapse that entire flow onto a phone: describe the page, let AI build it, edit it by talking to it, and publish it in one tap.

Technical Challenges

  • Targeted component editing: Long-press selection had to map a tapped region of rendered HTML back to the exact component so the AI could edit only that piece without regenerating the whole page.
  • Multi-component edits: Selecting several sections and applying one instruction across all of them meant batching edits into a single coherent AI request instead of independent, conflicting rewrites.
  • Static-viewer architecture: Instead of a full deploy for every page, a single viewer renders content dynamically from stored page data, so publishing is instant rather than pipeline-driven.
  • Cross-device sync: Auth via Google, Apple, or email, with pages and drafts stored in Supabase so a user's work follows them across devices.

Stack and Architecture

PageDub is a React Native mobile app. Generated pages and theme data are stored in Supabase, rendered through a static viewer, and published through a lightweight one-tap flow optimized for instant feedback instead of long deployment waits.

What's Next

The next step is strengthening template flexibility, theme composition, and the operational tooling around published pages as the builder ecosystem grows.

Interested in product systems that turn AI output into shipped software?

PageDub is a strong example of the kind of engineering work I enjoy: fast publish loops, clean constraints, and product-minded architecture.

Talk About This Project