Serposcope Proxy Setup: Format, Rotation, Pauses and Regional Checks

Serposcope takes proxies on its admin Proxies page, one per line, in a format with hash separators: http#IP#port#login#password for an HTTP proxy with credentials or socks#IP#port for SOCKS. After import you run the built-in check, delete the dead entries and then tune the Google networking options, where pauses and threads decide whether the tracker finishes quietly or drowns in captchas.

Plenty of tutorials show only the import window. This one also covers the details that trip people up: the IPv4-only rule, how Serposcope spreads queries across the pool, how to size pauses for your keyword count and how to make country and city checks reliable. One caveat up front: the project is no longer maintained, so check positions by hand now and then.

Serposcope in 2026

Serposcope is a free rank tracker you install on your own computer or server and use through a browser. The 2.x line is open source under the MIT licence; its GitHub repository was archived in December 2024 and the README describes the project as being in survival mode. The official site lists 3.5 as the latest release, with installers for Windows, Debian/Ubuntu and macOS, and states plainly that the program is no longer supported and may not work.

In practice that means two things. Existing installations keep running, but changes on Google's side are not followed by fixes. For example, in September 2025 Google stopped honouring the parameter that asked for 100 results per page, which affected many trackers that relied on it. Compare a handful of positions manually after major search updates.

Proxy Format: What Serposcope Accepts

The bulk import field understands four kinds of lines, according to the help text in the 2.x code:

Three rules matter more than they seem:

1. The address must be a numeric IPv4. The parser rejects hostnames with an invalid IPv4 error. If your provider gives you a gateway hostname, resolve it to its IP first and make sure that IP is stable, or ask for direct IP endpoints.
2. A single bad line cancels the whole import. Serposcope stops at the first invalid line and saves nothing, showing the line number in the error.
3. URL-style lines do not work. http://login:pass@host:port is a common habit from other tools, but here it fails as an invalid format.

The help text shows credentials only for HTTP proxies. The code reads login and password for SOCKS lines too, but if authenticated SOCKS gives you trouble, switch to HTTP or whitelist your server's IP with the provider.

Adding Proxies Step by Step

Step 1. Open the Proxies Page

Log in to the Serposcope web interface and open the admin settings. The Proxies item sits in the settings menu; some versions also show it as a tile in the settings overview.

Step 2. Paste the List

Click Add proxy, paste your lines into the bulk field and save. A success message confirms the import. If you see an error with a line number, fix that line and paste the whole list again.

Step 3. Run the Check

Start the proxy check from the same page. It runs as a background task; refresh the page to see the results as they arrive. Each proxy receives a status and, if it works, the outgoing IP it reported.

Step 4. Clean Up

Use Delete invalid proxy to drop everything that failed. Keeping dead entries wastes retries and slows every run.

Before importing a large pool, run a few addresses through our free proxy checker. It confirms the country of the exit IP, which matters for the regional setup described below.

How Serposcope Rotates Proxies

Serposcope does its own rotation. The help text for the threads setting says it never scrapes with the same IP in parallel, so each thread takes a different address from your list. Two practical consequences:

• Threads are capped by the pool. Five working proxies allow up to five parallel threads. More threads than addresses gives you nothing.
• Fixed addresses suit this model best. Since the tool already alternates between IPs, a list of stable exits keeps the behaviour predictable. A rotating gateway still works, but a query that fetches several pages may then load each page from a different IP. Our explanation of sticky and rotating proxies goes deeper into that trade-off.

Pauses, Threads and Retries

Open the Google settings and find the networking options:

The default five-second pause is far too short for any real project. A widely cited case study from 2017 tracked about 6,000 keywords with a pause of 60 to 120 seconds, five threads and nine proxies, and reported only a handful of captchas per run. Treat such numbers as a starting point, not a guarantee.

You can estimate the run time yourself: number of searches × average pause ÷ threads. For 300 keywords with an average pause of 90 seconds, one thread needs about 7.5 hours; five threads on five proxies cut that to roughly 1.5 hours. If the result does not fit into your daily window, add addresses rather than shortening the pause.

The settings also contain fields for third-party captcha-solving services. A well-paced setup should rarely need them; frequent captchas are a signal to slow down.

Regional and City-Level Checks

Search results depend on location, and Serposcope gives you three tools to pin it:

• Country sets Google's gl parameter for the search.
• Local sets the place you search from, as a country or city. The help text suggests Google's canonical location names, for example Paris,Ile-de-France,France.
• Custom parameters add anything else to the query string, such as hl=fr for the interface language.

Parameters alone are not always enough, because the IP address is also a location signal. For stable regional data, combine the right parameters with exits in the same country. Location-targeted residential proxies let you pick a country and often a city; just remember the IPv4 rule and check that the endpoint you use is a numeric address. For more on matching exits to search markets, see our guide to rank tracking proxies and the page on proxies for Google.

Troubleshooting

Invalid ipv4 at line N. You pasted a hostname or a malformed address. Use the numeric IP of the endpoint.

Invalid format at line N. The line has no hash separators, often because it was copied in host:port or URL form. Rewrite it as type#IP#port.

The check passes but searches return captchas. Lengthen the pause, lower threads and remove free or overused proxies. Datacenter ranges that many trackers share tend to hit challenges sooner.

Positions look wrong or empty. Check the country and local settings, then compare with a manual search from the same region. If the layout of Google results changed, the archived parser may no longer read it.