Клиентский проект29 Jan 2026
e-Proof - evidence capture for web pages
- Bun
- Express
- Puppeteer
- BullMQ
- Redis
- PostgreSQL
- Next.js 15
- WXT
- Astro
e-Proof records a copyright violation on a web page and packages it as one archive for claims work. The extension finds every image on the page, including carousels, modals and lazy-loaded ones; the user marks the infringing ones, and the backend produces a ZIP with an MHTML snapshot, a HAR traffic log, a PDF report with a QR code, WHOIS and traceroute data, and a SHA-256 hash per file, with progress streamed back over SSE. I took over an existing codebase and carried the whole cycle from there - markup and the client side, the server logic, and wiring every part into one system. The backend runs on Bun + Express with a pooled Puppeteer browser that recovers crashed instances and reaps zombie processes, BullMQ + Redis for the capture queue and Postgres for storage; all environment parsing lives in one module that fails fast on boot. The web app is Next.js 15 static export behind nginx, the extension is WXT + React with an alarm-based watchdog for when the stream drops, and the landing is Astro. About 400 automated test cases across backend, extension and end-to-end. Published in the Chrome Web Store and Firefox Add-ons.