Anatoly Skuba

I build scrapers that read the data a site already renders into its own pages, instead of parsing the DOM with CSS selectors. A redesign changes the markup — it does not change the payload — so these Actors keep working when selector-based scrapers quietly start returning nothing.

What that means in practice

• Measured failure rate, published openly. Every run writes a RUN_STATS record with the real numbers, so you can audit reliability yourself instead of trusting a claim
• No browser needed. Plain HTTP requests, so runs are fast and cheap
• Typed output. Numbers arrive as numbers, dates as ISO timestamps

No personal data

My Actors return company and agency names — business information — and never the personal contact details of individual agents or private sellers, even where the source page shows them. If you need personal data, these are not the right Actors.

Coverage

Dutch and Polish property portals, with more European markets in progress.