JOBFLOW

JobFlow is a local-first job search companion: dashboard, application tracking with a Kanban board, follow-up reminders, contacts, companies, email templates, analytics and automatic data-quality checks. Installable as an offline-capable PWA and deployed on Cloudflare Workers, with every piece of data kept in the browser's localStorage — full JSON/CSV backup and restore, no account, no server database.

PROJECT DETAILS

Objective

The goal was to replace a scattered spreadsheet-and-notes job search with a single local tool covering the whole pipeline — from sourcing offers to follow-ups — without ever requiring an account or sending personal data to a server.

Technologies used

The front end is TanStack Start (SSR) with React 19, TanStack Router and TanStack Query, styled with Tailwind CSS v4 and shadcn/ui components, and built with Vite via Nitro for deployment to Cloudflare Workers. All application data — applications, contacts, settings, email templates — lives in the browser's localStorage under versioned keys, exported and imported as backups with automatic migration from v1 to v4. A custom service worker (Workbox, injectManifest mode) caches assets and visited routes for offline use, with a dedicated offline fallback page and a non-intrusive update prompt. Recharts renders the analytics, and Playwright end-to-end tests (Python scripts) cover navigation and CSV import.

Challenges and solutions

The main challenge was making the app fully usable offline while keeping local data always available regardless of network state: solved with a custom injectManifest service worker that caches HTML per visited route (NetworkFirst strategy) without ever routing localStorage reads and writes through it. Backups also had to stay backward compatible as the data model evolved, which required a versioned JSON format with automatic migration from v1 through v4.