Skip to content
proxy testingresidential proxiesISP proxies

How to Test a Proxy Before You Pay

A repeatable proxy test: connectivity, exit country, sticky lifetime, and the fingerprint. No provider success rate required.

A provider's pool size, success rate, and average latency are not a test. They are a sentence on a homepage. This is a method you can run on ProxyLabs or on anyone else, with the free tools on this site, before you move real traffic.

ProxyLabs does not publish a pool size, a success rate, or a response time. The sheet at the end is blank on purpose. You fill it in from your target.

The four questions

A proxy purchase fails in four different ways, and a single "it works" check catches only the first.

  1. The client never connects, or it connects with the wrong protocol.
  2. The exit address is not in the country you asked for.
  3. A sticky session ends before the flow does, or it does not end when you think it does.
  4. The IP is fine and the browser still looks automated.

Write down the target URL, the country, and the session mode before you start. A result with no target attached cannot be compared with the next provider.

1. Does it connect?

Open the proxy tester. Paste up to 100 lines. It checks HTTP, HTTPS, SOCKS4, and SOCKS5, and it reports connectivity, response time, geolocation, and an anonymity read. No account is required.

Read the result as a connection check. A fast time to the tester's own URL is not the time to the site you actually need. A timeout here means the line, the protocol, or the gateway is wrong. It does not mean the target would have blocked you.

On ProxyLabs the protocols are not interchangeable:

ProductWhat you connect toProtocol
Residentialeu.proxylabs.app:8080 or us.proxylabs.app:8080HTTP
Mobilemobile-eu.proxylabs.app:5555 or mobile-us.proxylabs.app:5555HTTP, HTTPS, or SOCKS5, on that port
Dedicated ISPip:port:username:password from the dashboardHTTP first. SOCKS5 when the client asks for it

Americas residential traffic uses us.proxylabs.app. That list is us, ca, mx, br, ar, co, cl, and pe. Every other country code, including gb, uses eu.proxylabs.app. The mobile username is the one the dashboard generates. Do not rebuild it from the residential -ttl- pattern.

If a guide tells you to send residential traffic to a SOCKS5 port, stop and check which product the username belongs to.

2. Is the exit where you asked?

Take an address the tester accepted and open IP lookup. Check four things:

  • The public IP.
  • The country, against the country in the username.
  • The network name and ASN.
  • The proxy or VPN flag, and the WebRTC line. A WebRTC leak is a second address the page can see. It is not the same fact as "the proxy connected."

The residential username selects the country. The order is country, then region if you set one, then city if you set one, then -ttl-, then -session-. The United Kingdom code is gb. A username with uk is not the United Kingdom code this gateway uses.

text
customer-country-gb
customer-country-us-region-ny-city-newyork
customer-country-us-ttl-3600-session-abc123

Both country-us-city-newyork and country-us-region-ny-city-newyork are valid shapes. Use the city slug the dashboard offers. This page does not invent extra city names.

A country mismatch is a username or a gateway-host mistake. Fix that before you judge the target.

3. Does the sticky clock do what you think?

Rotating means you omit -ttl- and -session-. A new connection can leave from a new address.

Sticky means you set both. The lengths are 10, 30, 60, and 120 minutes, which are 600, 1800, 3600, and 7200 seconds. The clock starts when the session starts. Using the same session id later does not add more time. When you need another address, generate a new session id. When you need the same address tomorrow, that is a dedicated ISP proxy, not a longer sticky value.

Mobile sticky lengths are different: 1, 10, 30, 60, 120, 360, or 1,440 minutes. 1,440 minutes is 24 hours. Copy that username from the dashboard.

A 10-minute residential check you can actually wait for:

python
import requests

# 600 seconds is 10 minutes. The clock starts at the first use of this session id.
proxy = "http://customer-country-gb-ttl-600-session-abc123:[email protected]:8080"
session = requests.Session()
session.proxies = {"http": proxy, "https": proxy}

first = session.get("https://httpbin.org/ip", timeout=30).json()
second = session.get("https://httpbin.org/ip", timeout=30).json()
print("inside the window", first, second)

Inside the window, those two reads should show the same address. After the window, the same session id is not a promise of a new 10 minutes. Note the clock time next to each read. "It changed" with no timestamp is not a failed sticky session. It may be a session that already ended.

The longer account of when to rotate and when to stick is rotating vs sticky proxies. The username map is how to connect.

4. The fingerprint is a separate test

A residential address does not change navigator.webdriver, the canvas, WebGL, fonts, or audio. Those are the browser. Open the fingerprint test in the browser you will actually ship, through the proxy if the flow is in that browser.

If the IP lookup is clean and the fingerprint test shows an automation marker, another proxy will not hide that marker. The Playwright notes for six of those markers are in the Playwright proxy setup.

Read the price before the adjective

Compare the invoice, not the adjective "premium."

Residential bandwidth lists at £2.50/GB, is £1.00/GB from 5,000 GB, and is £0.95/GB from 10,000 GB. Bought data does not expire. There is no monthly minimum. The trial is 0.5 GB for £1. It is not a free test.

A dedicated ISP proxy is one address for the month, unmetered, in the countries on the ISP page. The price is £1.50 per IP per month, down to £1.18 at 2,000+ IPs. Japan and Hong Kong are £2.00 per IP per month, down to £1.68 at 2,000+. There is no ISP trial.

Mobile bandwidth lists at £3.00/GB and is £1.50/GB from 10,000 GB. The product page is mobile proxies.

If a comparison quotes a success rate, ask for the target, the client, the country, and the date. If those four are missing, the number is not a result.

For response time, measure the median and the slow tail on your target. A single average against a fast echo server hides the slow requests. The measurement notes are in how to test proxy speed. That article does not publish a ProxyLabs latency, and this one does not either.

A sheet you can copy

Use one row per provider. Leave the cells empty until you have run the check.

ProxyLabsOther provider
Target URL
Country asked for
Mode: rotating, sticky, or dedicated
Connected, and which protocol
Exit country matches
Same IP on two reads inside the sticky window
Time the sticky window ended
WebRTC showed a second address
Fingerprint markers you care about
Page result on the target
Price for the amount you will actually buy

Run the sheet twice if the first hour looked unusual. One afternoon on httpbin.org is a connection test. The decision is the row for your target.

When the residential row is the one you want, the buy page is residential proxies. When the address has to survive until tomorrow, the buy page is ISP proxies. When the check is a carrier network, the buy page is mobile proxies.

Useful guide? Share it with your team.

Share
← Back to all articlesBack to top ↑
All articles ↗
residential proxies

Residential Proxies in the US: Providers, Pricing, and Setup

How to target a US residential address: the Americas gateway, the city slug from the dashboard, and how to check the exit before a crawl.

7 min read
residential proxies

Why your residential proxy still gets blocked on Akamai and Cloudflare sites

A residential exit can be clean and still meet a block page. The address, the TLS handshake, and the challenge script are three separate checks.

6 min read
dedicated proxies

Dedicated Proxies: What the Address Includes

What a dedicated proxy includes on ProxyLabs: one ISP address for the month, the countries, the price, and what the plan leaves out.

5 min read