Work · 2026

Social archive tooling

A Laravel and Filament dashboard, backed by a Playwright scraper, for archiving a brand's social media with a human review step.

Side project · 2 min read · event #000036

  • PHP 8.4
  • Laravel 13
  • Filament 3
  • Node.js
  • Playwright
  • Cloudflare R2
  • Coolify

The problem

A brand’s social media history lives on platforms that can change or remove it at any time. I wanted my own copy of the posts and their media for the shop’s account: something I could search, review and keep, instead of trusting that the platforms would hold it for me.

What I built

There are two parts, joined by a queue of work.

The collector. A Node.js scraper using Playwright drives a real browser to visit the account and collect posts and their media. A browser is slower than an API, but it sees what a visitor sees, which is what I wanted to keep. Media files can be large, so they are streamed straight to Cloudflare R2 rather than being held in memory or written to the server’s disk first.

The dashboard. A Laravel 13 application on PHP 8.4, with the admin built in Filament 3. It lists what has been collected and lets me work through it.

Human review. Nothing goes into the archive on the scraper’s say-so alone. Collected items arrive in a review queue and I approve, reject or annotate them. For an archive that is meant to be kept for years, I would rather spend a few minutes of attention than discover later that it contains duplicates or the wrong account’s posts.

The whole thing is deployed through Coolify on the same server as my other projects, so it follows the same deploy and backup habits.

Choices worth explaining

I kept the scraper in Node because Playwright is best supported there, and kept everything else in Laravel, which I can maintain fastest. The two parts only meet at the queue and the storage bucket, so either could be replaced without touching the other. Storing media in R2 keeps the application server small and the archive independent of the machine it runs on.

What I’d do next

Add checksums for stored media so that the archive can be verified later, and make the review queue show a side-by-side comparison when a post looks like a near-duplicate.

← Back to work