How to set up rotating proxies in ScrapeBox

ScrapeBox is a Windows SEO tool built around a search-engine harvester: you give it keywords and footprints, it queries Google, Bing and about 30 other engines, and it collects the result URLs. Around that sit a keyword scraper, a proxy manager and dozens of addons for checking links, metadata, indexing and more. It runs on Windows desktops and Windows servers alike.

Almost everything ScrapeBox does sends many queries to the same few sites, so it needs proxies. Without them, Google starts returning CAPTCHAs and empty pages within minutes. With a rotating gateway, each connection leaves from a different IP, so the load spreads over a large pool instead of a short list of static IPs that get burned one by one.

Most trouble with rotating proxies in ScrapeBox comes from habits formed with proxy lists: running the proxy test and deleting “failed” gateways, setting 100+ harvester connections against Google, and forgetting that addons running at the same time share the same threads. This guide walks through the setup and the settings that matter.

ScrapeBox’s Proxy Manager can test proxies for speed, anonymity and Google access, then filter or remove the ones that fail. That checker isn’t compatible with Storm rotating gateways. It may mark working gateways as failed or slow, and if you then use “remove failed” or filter by speed, you delete proxies that work perfectly in the harvester.

The proxies themselves work fine in ScrapeBox; only the checker doesn’t match how a gateway behaves. So:

• Don’t run the Proxy Manager test on gateway lists, and don’t filter them by the test result.
• If you enabled auto-removal of dead proxies during harvesting, turn it off for gateways. A failed query usually means one IP was rate-limited on that request, and the next connection through the same gateway goes out from a different IP.
• Verify with a browser instead: set one gateway as the HTTP proxy in Firefox or Chrome, open an IP-check page, and refresh. A new IP on each visit means the gateway and your authorization are fine.

A rotating plan’s thread count is the number of connections it accepts at the same moment. ScrapeBox lets you set hundreds of harvester connections, but Google reacts to how fast queries arrive, not just to how many IPs send them. Running the whole plan against Google floods it with queries, and the share that comes back as CAPTCHAs or empty pages climbs fast.

Keeping Google and other search engines to at most a quarter of your threads holds the query rate where most of them return results. In practice:

• 40 threads: 10 harvester connections.
• 80 threads: 20 connections.
• 150 threads: 37 connections.
• 200 threads: 50 connections.

Need results faster? Move up a plan rather than raising connections on the same one. Engines other than Google, and non-search tasks, can use more of the plan, as long as the total stays inside it. The Detailed Harvester also lets you add a delay between queries, which helps on long footprint runs.

The plan’s limit applies to everything connecting through your gateways at once, from every program on every authorized machine. ScrapeBox addons such as the Alive Checker, Link Extractor, Page Authority and Google Meta Scraper, plus the built-in Keyword Scraper, each have their own connection settings, and they all draw from the same pool of threads.

• Running the harvester and an addon together? Add their connection counts. A 10-connection harvest plus a 50-connection Alive Checker run is 60, which is over a 40-thread plan.
• The Keyword Scraper pulls from search suggestion sources such as Google Suggest and Bing, so count it inside your 25% search-engine share.
• A browser tab you open for checks uses around 10 threads while it loads.
• Set each addon’s connections before starting it rather than leaving defaults, which are often far higher than a small plan allows.

Going over the limit doesn’t make the job faster. Extra connections fail, and ScrapeBox shows them as proxy errors, which then look like a broken proxy.

• Main (new IP per connection): search-engine harvesting, keyword scraping, link and index checks. This is the gateway for nearly all ScrapeBox work.
• 15-minute: tasks that must keep one IP across several requests to the same site, such as stepping through pages of one domain.
• 3-minute: meant for account registration, social sites and browsing, USA only. Not for search engines, so keep it out of your harvester list.

Locations work the same way: USA, EU or USA+EU for region-specific checks, and Worldwide for broad scraping. The gateway IP:PORTs are fixed, so once they’re in a saved ScrapeBox proxy file you never need to update the list; the rotation happens on our side.

Many people run ScrapeBox around the clock on a Windows VPS. That works well with IP authorization: authorize the VPS IP, not the PC you connect from with Remote Desktop. If you run ScrapeBox on both, authorize both IPs and split the threads between them. If your home IP changes, authorize a free dynamic DNS hostname such as No-IP instead of a raw IP. For other SEO crawlers on the same plan, see the Screaming Frog guide.