Summary of an article published on The Web Scraping Club on 2025-05-16 (for paid subscribers). Read the full article
In 2025, the hardest websites make an anti-detect browser hard to skip in a scraping stack, and Camoufox is the open-source one with a good track record. Browser-as-a-service vendors take the infrastructure off your plate, but running Camoufox as a server inside a container and scaling it as operations grow is a hot topic in our Discord community.
I build a working Camoufox container, write a test scraper that connects to it, and go through the pros and cons of the approach. Before that, I cover containers and Docker for scraping: how they differ from virtual machines, why isolation, resource limits and reproducibility matter, and why a classic crawler gets little from horizontal scaling while a scraper exposed as an API suits it well. The Camoufox server is the second use case, with a Playwright scraper connecting to it.
What’s inside
- Why spawning more identical containers replicates a saturated scraper instead of fixing it, and what to do about it.
- How versioned images let you roll back to a known working environment when something breaks.
- Why frequent code changes that force image rebuilds add overhead for a traditional scraper deployment.
- A short primer on the Dockerfile instructions, the image build and the container run, used as the base for the Camoufox server.
In the full article: Dockerfile for the Camoufox server, the Playwright test scraper, the connection setup and the full list of pros and cons, plus the scripts in the 83.CAMOUFOX-SERVER folder of the GitHub repository.