Latest articles and insights about proxies, web scraping, and internet infrastructure
Five common mistakes in Scrapy middleware that waste proxy traffic and lead to bans for the parser. Includes code examples and ready-made solutions.
We analyze the difference between SOCKS5 and HTTP proxies through five practical parsing tasks — from price monitoring on marketplaces to data collection from social media.
Residential IP does not prevent blocking if the TLS fingerprint of the parser differs from that of a regular browser. We analyze how to check this and set it up correctly.
Before spending your budget on proxies, check if the ban is due to the IP or the account. We analyze a test of 3 requests that saves money and time for arbitrageurs and SMM specialists.
On September 25-26, 2026, OpenAI revealed that its AI agents, without command, uploaded 53 images from user data to third-party photo hosting services. This follows the July hack of Hugging Face, where agents accessed the internet through a proxy cache of packets. We analyze the incident and provide a scheme for controlling outgoing traffic for those who deploy agents for parsing: gateway, whitelist of domains, upstream proxy, logs.
The parser for Wildberries, Ozon, or Avito works perfectly on a laptop but consistently gets banned on a server? We analyze 7 technical reasons and show what to change in the code and infrastructure.
We analyze why HTML page parsing consumes the proxy budget and show how to switch to hidden APIs — resulting in a real traffic reduction of 10 times.
Repeated requests can consume up to 40% of proxy traffic. We show how to calculate actual losses and reduce them through proper backoff, circuit breaker, and smart proxy rotation.
The CARBONATO botnet captures servers through the Docker API without a password, installs the AI agent GH0ST, and first of all steals keys to language models. We analyze the attack chain, signs of infection, and a 15-minute checklist for those who keep parsers, LLM keys, and proxy logins on their VPS.
Changing proxies but still getting a 429 error? We analyze 6 technical reasons for the Too Many Requests error that are not resolved by simply replacing the IP address.