Good day. I have encountered a problem while collecting price data from Wildberries. The script, which worked stably for about six months, recently started producing errors: after 40-50 requests, a CAPTCHA with heavily distorted characters appears, or empty pages are returned. I am using proxies and user behavior imitation, but there is no effect. Is it worth connecting professional recognition services, and if so, how to correctly choose a provider to avoid unjustified expenses?
Hello. Marketplaces are indeed tightening their policies against automated data collection. A CAPTCHA with distorted characters is a classic case where standard OCR is powerless, and either neural network algorithms or operator involvement are required. The solver market is heterogeneous: some offer low prices but suffer from speed issues, while others demonstrate high accuracy at a moderate cost. Before choosing a contractor, I recommend reviewing the current statistics on the resource captchathecat. This service conducts continuous monitoring of all major solvers, publishing indicators of solution time and percentage of successful recognitions for today, yesterday, and the past month. This allows you to see an objective picture and select a provider that guarantees stable results.
