X (Twitter) Proxy

X detects heavy logged-out browsing and automated search requests within minutes and rate-limits traffic arriving from a single IP; if the pattern continues, a temporary CAPTCHA or an access restriction follows. Spreading requests across a large residential or mobile IP pool keeps the per-IP request rate inside what the platform tolerates, so your scraping job keeps running without interruption.

There is no fixed number; it depends on account age, behavior patterns and how sensitive the platform's detection is at that moment. The general rule is to keep every account you do not want linked to the others on its own residential or mobile IP, in sticky mode. Several accounts logging in from the same IP at the same time can lead the platform to connect them and suspend them in one sweep.

Controlled concurrency means capping, per IP, how many requests go out at once and how fast. Firing dozens of threads through a single exit or a narrow IP pool runs into rate limits and bans in short order. Spreading requests across a large pool and lowering the request rate per thread lets you collect the same total volume of data at a far lower detection risk.

Start by lowering your request rate and your number of concurrent connections, and add backoff and retry logic for 429 responses. If the problem persists, spread traffic across a large residential or mobile pool instead of a single IP, and use sticky sessions on account-based work so the IP does not keep changing — frequent IP switching is itself read as suspicious behavior.

Residential proxies give you the better cost-performance balance on high-volume general search and timeline scraping. Mobile proxies carry the highest IP reputation and are the pick for account-based work — multi-account management above all — where the platform treats mobile network traffic more leniently; in return, the cost per GB is higher.

Official API access to X comes with tiered pricing and limited request quotas; on high-volume or exploratory needs (real-time trend tracking, for example) those quotas run short fast. Collecting data from the public web interface through a proxy lets you work at a more flexible volume and pace without being bound by those quota limits; complying with the target site's terms of use remains your responsibility.

Some posts and trend lists return different results by region; the country or city parameter you append to your username narrows the pool so you search only through IPs exiting from that region, letting you test regional trends and post visibility exactly as they appear there.

You define the backconnect endpoint (host:port and username:password) as the proxy server in the browser launch parameters; there is no need to keep a separate IP list or rotation logic in your code, a single address is enough. In sticky mode you can append a duration parameter to the username to keep your session identity fixed.