Keeping Claz useful for people, not bots
Claz is built for people – someone selling a couch, looking for an apartment, replying to an ad nearby. Lately, a growing share of our traffic hasn’t been people at all. It’s been bots.
Why now
Section titled “Why now”AI and large language models have made large-scale scraping cheap and easy. The result is more automated traffic, often not gentle. That worries us for three reasons:
– Abuse of real users’ data. Listings posted in good faith can be harvested in bulk and reused for scams – fake offers, phishing, impersonation.
– Reliability. A flood of automated requests can slow Claz down for the people actually here to buy and sell.
– Fairness. Scrapers take everything and give nothing back.
What we did
Section titled “What we did”The hard part isn’t blocking traffic – it’s blocking the right traffic, without annoying real users. So, at a high level:
– Smarter traffic limits, so no single source can hammer the site or vacuum pages at machine speed.
– Abuse detection that separates automated harvesting from normal human use.
– Verification only when needed – a quick check shown when traffic looks suspicious, invisible in normal use.
What this means for users
Section titled “What this means for users”Probably nothing. Browsing, searching, and posting should feel the same. On rare occasions a user might be asked to confirm they’re human – that’s just the system being careful, and it should pass quickly.
Tell us if we got it wrong
Section titled “Tell us if we got it wrong”We’d rather hear that a measure is in your way than have you quietly leave. If anything slows you down or blocks something legitimate, please tell us – we’ll tune it. Getting this balance right is ongoing work, and your feedback is the best signal we have.
