Skip to content

2025 · Full-Stack / Systems

22 Autonomous Apify Scraper Fleet

A fleet of 22 web scrapers orchestrated end-to-end with an AI agent supervisor inside a 900MB sandbox.

22 Scrapers · Agent-Orchestrated Build

Apify scraper fleet dashboard
Autonomous scraper pipeline monitoring and dataset delivery.

The problem

Building and maintaining dozens of distinct scrapers manually against evolving website layouts and anti-bot protections is too slow. Hand-writing each crawler one by one was never going to scale.

The constraint

The scrapers had to operate inside a strict 900MB RAM server slice without memory leaks, cycle proxies to avoid rate limits, and abort cleanly without billing customers if a target website layout shifted.

What I built

An autonomous supervisor loop that mined search demand (targeting queries with demand_runs_30d >= 400), synthesized TypeScript / Crawlee scrapers, executed pre-flight dry runs, configured Pay-Per-Event pricing, and deployed to Apify cloud containers.

The outcome

22 production regulatory and compliance scrapers shipped and monitored, with automated canary tests preventing broken runs from ever billing users.

Links