Our proxy knowledge base

We sell two types of proxies: datacenter proxies and ISP proxies (static residential). We don't offer peer-to-peer residential or mobile proxies.

Datacenter proxies use IP addresses that belong to data centers and hosting providers, so an IP lookup shows the address as datacenter-owned. They're the most affordable option, with high speeds, stable performance, and unlimited bandwidth. They fit most use cases: web scraping, SEO monitoring, ad verification, automation, and API work. You can buy them as packages starting at 100 IPs, as dedicated proxies starting at 1 IP, or as rotating proxies.

ISP proxies (static residential) use IP addresses that belong to consumer internet service providers but run on our servers in a data center. IP databases associate them with an ISP rather than a hosting company, which gives them residential-level trust, while speed and stability stay at datacenter level. The IP stays the same for your entire plan term. Learn more on the ISP proxies page.

In short: choose datacenter proxies when volume and price matter most, and ISP proxies when the target website needs the IP to look like a regular ISP address. Each type is covered in detail in “What are datacenter proxies?” and “What are ISP proxies?”. For targets with strict anti-fraud systems, see “Do you have proxies for betting, payment services, or casinos?”.

Keep in mind that datacenter proxy packages are shared: up to 3–5 customers can use the same IP at a time. Lists aren't prepared in advance — each package is built automatically from our entire pool, so different customers get different sets of IPs. That's how we keep prices well below the market. If you need an IP reserved exclusively for you, choose dedicated proxies. For possible overlap between packages, see “If I buy two identical packages…”.

Datacenter proxies are IP addresses hosted in data centers; no real home users sit behind these IPs — they aren't tied to any household connection. They're a good fit for most business use cases, with low cost, high stability, and unlimited bandwidth. That said, we don't recommend them for use cases where the target scrutinizes every user in detail (anti-fraud checks) — for example, signing up at betting sites, registering with banks and payment services, claiming cash bonuses across various platforms, and the like. For everything else, datacenter proxies offer an unbeatable price-to-performance ratio.

ISP (Internet Service Provider) proxies are static residential IP addresses that officially belong to internet service providers but are hosted on our servers in data centers. That gives them the best of both worlds: the trust of residential addresses (sites see an ordinary home connection from an ISP) plus the speed and stability of datacenter proxies.

Their key traits: they're static (the IP doesn't change during the paid period), fast, and come with unlimited bandwidth, and they enjoy higher trust from websites than regular datacenter proxies. They're a good fit wherever you need both IP trust and performance at once. Learn more on our ISP proxies page.

Datacenter proxies are hosted in data centers — a check will show the datacenter's IP. Residential proxies use the IPs of real home users.

ISP proxies are static residential IPs that belong to internet service providers but are hosted on our servers. They combine the trust of residential addresses (sites treat them as an ordinary home connection) with the speed and stability of datacenter proxies. That makes them a good fit wherever you need both IP trust and high performance at once. Learn more on our ISP proxies page.

Dedicated proxies are IP addresses that belong to you alone for the entire paid period. You don't share the bandwidth, connection limits, or IP reputation with anyone else.

In practice, that means you get the full bandwidth and all the connections in the package; the address's reputation depends only on your activity, not other users'; and you get higher stability and predictability. Dedicated proxies cost more than our shared packages, but they give you maximum control and suit use cases that need a clean personal IP — account management, social media and messaging apps, and access to streaming services. You can get them on our dedicated proxies page.

With a regular package you get a fixed list of IP addresses and decide which one to use yourself. The list only changes when you want it to (see “I bought a proxy package, but I want to refresh the IP list”).

Rotating (backconnect) proxies work differently: you connect to a single entry point, and on each request the system automatically hands you a new IP from the shared pool. You don't see or manage specific addresses — rotation happens on our end. This is handy for web scraping and working with third-party APIs, where automatic IP rotation matters. How requests are counted on rotating proxies is covered in “What are API credits in rotating proxies?”.

Our proxies cover a wide range of use cases, and they're a great fit for most common scenarios. First and foremost they're built for professionals who need large-scale data collection, but they work well for everyday use too.

The most common use cases: web scraping and data collection (marketplaces, search results, price monitoring), SEO monitoring across regions, ad verification, multi-account management on social media and messaging apps, access to regional and media content, streaming, automation via bots and APIs, testing, and getting around geo-restrictions.

For everyday use cases that need a stable personal IP, we also offer dedicated proxies. We only advise against datacenter proxies for niches with strict anti-fraud (bookmakers, payment services, casinos) — see “Do you have proxies for betting, payment services, or casinos?”.

You can configure an available plan for a non-refundable 48-hour trial, or use the 24-hour refund window on a standard-term service.

If a package says 100 IPs, you get 100 different IP addresses. Different ports on the same IP don't count as separate proxies. So “100 IPs” always means 100 unique addresses, not a single address opened on several ports.

No, we provide IPv4 only — it's universal and works for virtually every use case.

We avoid IPv6 on purpose: the protocol has real downsides. Many sites and services still don't support it, or treat such addresses with suspicion. IPv6 is also handed out to providers in huge blocks, so it's easy to recognize an address as part of a “batch” and block whole subnets at once. The result is noticeably lower stability and compatibility, whereas IPv4 delivers a predictable, clean result.

These niches (bookmakers, payment services, casinos) run strict anti-fraud systems that scrutinize every user in detail. Datacenter IPs are relatively easy to identify, so they're usually a poor fit for this kind of task. What you typically need here are residential or mobile proxies, which we don't offer.

If you don't already have experience running datacenter proxies on such a target, configure a non-refundable 48-hour trial or use the 24-hour refund window on a standard-term service.

Proxies are delivered automatically right after payment. We don't send credentials by email: the service, with its username, password, and IP list, appears in your Dashboard under “Products/Services.”

If the service isn't activated automatically within 10 minutes of payment, open a ticket in your Dashboard, and we'll check the payment and activate the service.

Yes. Configure any available plan with the locations, quantity, authentication method, and add-ons you need, then pay for a 48-hour trial. Trial payments are non-refundable. If it works for you, extend the same service for a standard term.

You can pay in many ways:

Cards and wallets: Stripe, Visa, Mastercard, American Express, Apple Pay;
Cryptocurrency: USDT (TRC20/ERC20), USDC, Bitcoin, Ethereum, Litecoin;
Other methods: WebMoney.

The current list of available methods is always shown on the invoice payment page in your Dashboard.

You can get a full refund if the proxies aren't a fit — just request it within 24 hours of purchase.

Refunds are issued using the same method you used to pay for the service. To request one, just contact support. You can find the detailed refund rules here.

After the first 24 hours, you can cancel an order and have any unused funds credited to your account balance. You can cancel an order once every three days from your Dashboard, without contacting support.

You can renew a service from the “Products/Services” section of your Dashboard, for any available term. Auto-renewal is available too: payment is charged automatically from your account balance, or — if you've paid by card before — from your saved card.

When you renew the same service, your IP addresses stay the same — the list doesn't change unless you refresh (replace) the IP list yourself (see “I bought a proxy package, but I want to refresh the IP list”). This lets you keep working without reconfiguring whitelisting or your software.

Yes. The longer the term, the lower the monthly price:

3 months — 10% off;
6 months — 15% off;
12 months — 20% off.

The discount applies automatically when you select a term in the calculator on our website or in the checkout configurator in your Dashboard. No promo code needed, and you see the final price before you pay.

For example, a $19/month plan would cost $228 for 12 months without a discount. With 20% off, you pay $182.40, or about $15.20/month.

Not sure the proxies are a fit yet? Start with a 48-hour trial or a shorter term, then renew for a longer period.

You don't need to place a new order — you can add IPs right to your active service:

1. Open your active service in the Dashboard.
2. Go to “Management” and click “Buy more IPs.”
3. Choose a country and a quantity. You can add more IPs in a country you already have or add IPs from a different country.
4. Review the amount and confirm payment from your balance.

Pricing. You pay only for the days left in your current billing period — the charge is prorated. Before you pay, you'll see both the amount charged now and the new renewal price for the service.

The new IPs are added to the same service, so you don't need to set up IP whitelisting again.

The standard minimum term is 30 days, and you can renew for longer periods. The only shorter option is a paid 48-hour trial, which lets you test the proxies with your actual use case. You can also request a refund within 24 hours of payment if the IPs aren't a fit for your use case for any reason.

For the full term you've paid for. Our proxies don't “die” during the paid period: if your package runs for 30 days, the proxies will work the full 30 days without replacement. The same goes for a 365-day term.

All our proxies use mandatory IP whitelisting — they only work from the IP address you whitelist. This isn't just a technical requirement; it's an important feature that gives you real advantages.

IP whitelisting lets us give you truly unlimited bandwidth and full speed on every proxy, so you can use them to their full potential without worrying about limits. It also maximizes security, since only you can use your proxies. Even if your username and password leak to a third party, they won't be able to use the proxies from another device, which keeps your data and your work protected.

To set it up, find the real public IP of the computer or remote server you'll run the proxies from (disable any VPN before checking). You can do this at showmyip.com. Write down the IP you get.

Next, in your Dashboard (https://papaproxy.net/account_new/login?locale=en), open your active service.
On the service page you'll see the username and password for proxy authentication, plus a field where you enter the IP you got from showmyip.com. Enter your IP and click “Set.” The new IP starts working within 15 seconds, after which your proxies are ready to use.

Within 15 seconds of saving the IP under “Management” → “IP Whitelist” on your active service page. There's no need to wait longer or contact support.

If the proxies still don't respond after a minute, check the following:
1. You added the public IP of the device you work from. You can find it at showmyip.com — disable any VPN first.
2. Your software actually connects to the internet from that IP. For example, a server or VPS usually has a different IP than your computer.
3. You're testing from that same device, not through an online proxy checker — see “A proxy checker shows that my proxies don't work”.

For step-by-step whitelisting setup, see “What is IP whitelisting and why is it needed?”.

In the “Products/Services” section of your Dashboard, click your active service. The username and password for proxy authentication are shown just above the field where you whitelist your IP.

Yes. You can change the proxy username and password at any time on your active service page in the Dashboard. Change either one, or both at once.

Don't confuse these with your Dashboard login. They're a separate set of credentials your software uses to authenticate with the proxies. Your account password is changed in your profile settings.

When to change them:
— your credentials may have been exposed (for example, in a screenshot or a shared config file);
— you gave access to an employee or contractor and want to revoke it;
— you want a recognizable username to tell services apart in your software.

Keep in mind:
— once you change the password, the old one stops working immediately;
— the new password is shown only once, so copy it right away;
— update the credentials in every application that uses these proxies, or they'll get an authentication error (status code 407).

If you connect only through IP whitelisting (ports 8085, 8444, 1085, and 1087), your proxies don't use a username and password, so changing them won't affect your connection.

We support the HTTP, HTTPS (HTTP/2), SOCKS5, and SOCKS4 protocols — all available at the same time, differing only by port. Every proxy list we issue already includes the port: just download your list from your Dashboard, and the port is shown next to each IP.
Ports by protocol and authentication method:

HTTPS also has a backup port, 993, which you can use in addition to 8444.
* SOCKS4 has no username/password authentication in the protocol itself, so it's available via IP whitelisting only.
A note for developers: SOCKS4 and SOCKS5 run on separate ports.

Most providers combine SOCKS4 and SOCKS5 on a single port. As a developer-focused service, we deliberately keep them apart — SOCKS5 on 1085, SOCKS4 on 1087. This matters because some software is built on the outdated SOCKS4 protocol but labels it “SOCKS5.” In that case our SOCKS5 port may not work with it — point such software to our SOCKS4 port (1087) instead.

Every proxy works over several protocols at once — HTTP, HTTPS (HTTP/2), and SOCKS5 (plus SOCKS4). There's nothing to switch: they're all available at the same time and differ only by port.

HTTP — the standard protocol for web traffic: loading sites, scraping pages, working through a browser. A universal choice for most use cases.
HTTPS (HTTP/2) — the same, but over an encrypted channel with full CONNECT support, so your traffic stays encrypted on the leg between your device and our service. Choose it when connection privacy matters most.
SOCKS5 — the most flexible proxy protocol: it works at a lower level and passes any kind of traffic, not just HTTP, which suits non-standard protocols, desktop applications, and high-throughput workloads.

If you're unsure, start with HTTP (port 8080 with username/password authentication) or use HTTPS for an encrypted channel. The full port table is in “What ports and protocols do you support?”.

Yes. UDP is a paid add-on available for all plans except rotating proxies: datacenter proxy packages, dedicated proxies, and ISP proxies.

Do you need it? Browsers, scrapers, and most applications use TCP, and all TCP traffic works without UDP. UDP is used by voice and video calls, online gaming, streaming, some messaging apps, and certain network tools. If your software requires UDP or reports that the proxy doesn't support it, add the option. UDP traffic runs over SOCKS5.

How to add it:
— for a new order, select the UDP option in the checkout configurator, and it's enabled on every IP in the package;
— for an active plan, add it on any day of the term from your active service page.

Pricing. The add-on costs 50% of the base plan price. If you add it partway through the term, the charge is prorated for the remaining days. If you cancel, unused days are prorated as well.

The add-on renews together with your plan, and IPs you receive after a proxy list refresh support UDP too. Learn more on the SOCKS5 UDP proxy page.

Quick clarification: HTTP, HTTPS (HTTP/2), and SOCKS5 are protocols, not formats, and they're all available at the same time (nothing to switch).

The proxy list itself can be downloaded from your Dashboard in TXT, CSV, and JSON — with the right ports for username/password or IP (whitelist) authentication. There's also a format builder: you set the field order, separators, and line template to match your software.

On top of that, there's a link to import the whole list straight into your program, and a separate link that returns one random proxy from your package on each request.

Yes, you can. If your ISP changes your IP address within the same subnet, IP whitelisting in your Dashboard will keep working without constant updates.

Simply enter your current IP in the settings, use username/password authentication, and the system will automatically determine the allowed range in which your proxies will remain accessible. For HTTP proxies with authentication, use port 8080. For SOCKS5 proxies with authentication, use port 1080.

How does it work?
If your IP is, say, 85.249.3.145, the system automatically expands the range to 85.249.0.0–85.249.7.255 (a /21 subnet mask). Even if your IP changes within this range, the proxies will keep working without interruption.
Additionally, you can check your range using a CIDR calculator, for example here, but you don't need to whitelist the entire range manually — the system will handle it automatically.

Yes. You can whitelist one package to multiple external IPs and use it on several devices or servers at the same time. The first five IPs are included at no additional charge. Each additional IP whitelisting rule costs $1/month.

The concurrent connection limit (more about limits) is shared across all whitelisted IPs, and you control how it's split from your Dashboard — see “How do I split the connection limit across multiple whitelisted IPs?”.

You can add more than five IPs as needed at that price.

No, whitelisting is mandatory — and it works in your favor: it's exactly what lets us give you truly unlimited bandwidth and full speed, while keeping your proxies safe from outsiders. Even if your username and password leak, no one can use the proxies from another device (more in “What is IP whitelisting and why is it needed?”).

If you have a dynamic IP, the proxies work with username/password authentication within a /21 mask — no extra setup needed. If you plan to use one large package across different servers, whitelist it to multiple external IPs. The first five are included at no additional charge, and each additional IP whitelisting rule costs $1/month.

Our proxies are universal and work anywhere you can specify an HTTP, HTTPS (HTTP/2), or SOCKS proxy: on Windows, macOS, and Linux, on servers (VPS/VDS), in browsers, antidetect browsers, scrapers, bots, and most desktop applications.

The one requirement: your traffic must originate from the device (computer or server) whose IP is whitelisted to the package in your Dashboard. Mobile devices and apps can use the proxies too, as long as they go online from the whitelisted IP. More on whitelisting in “What is IP whitelisting and why is it needed?”.

Every server has an 80 Gbps uplink. We don't apply artificial caps to proxy speed, which reaches up to 500 Mbps per proxy. You can read more about speed here — https://papaproxy.net/fast-proxy.php. Please also stay within the bandwidth fair use guidelines for your plan — see our Terms of Use.

No, bandwidth is genuinely unlimited — there are no data caps, and your speed doesn't drop no matter how much you transfer. We only ask that you stay within the bandwidth fair use guidelines for your plan (see the question about speed above).

A connection is an open channel between your software and the proxy. Each time your browser, scraper, or script sends a request through a proxy, it opens a connection. The connection counts as active while the request is processed and stops counting once the response arrives and the connection closes. Concurrent connections are how many of these channels are open at the same moment — that's the number your plan limits (see “How many connections am I allowed?”).

One page isn't one connection. A web page is made of many resources: HTML, images, stylesheets, scripts, fonts, API calls, and analytics trackers. The browser loads them in parallel, and each resource can use its own connection. Rough guidelines:
— a simple API request or a single HTML fetch without a browser — 1 connection;
— a page in a browser without JavaScript — 2–3 connections;
— an online store page in a regular or headless browser — 10 to 40 connections;
— heavy pages with ads, trackers, and analytics — even more.

Threads and connections aren't the same thing. A thread is a parallel task in your software (for example, 50 threads in a scraper). Each thread can open several connections at once, so 50 threads in browser mode can easily generate several hundred concurrent connections.

How to estimate your usage before you buy:
1. Open the target website in Chrome or Firefox, press F12, and switch to the Network tab. Reload the page and see how many requests it makes. You can also get a visual request map with RequestMap.
2. Multiply that number by the number of threads in your software.
3. Add a 20–30% margin for peak load.

Close your connections. Configure your software to close each connection when its task finishes. Stuck threads and connections your software keeps open “just in case” still count toward your limit even without traffic. That can push you to the limit and into Slow Mode while very little real work is happening. If your connection stats don't match what you expect, ask your software developer to check how it handles connections.

For more examples and ways to reduce your connection count, read “Understanding the concurrent connection limit.”

It depends on your package. The limit is aggregate — it applies to the whole service, not to each IP.

Packages up to 999 IPs: 1,500 concurrent connections.
From 1,000 to 4,999 IPs: 3× the number of IPs in your package.
From 5,000 to 14,999 IPs: 2× the number of IPs in your package.
Packages with 15,000+ IPs: the limit equals the number of IPs in the package.

You can always see your current connection stats in your Dashboard — see “Where can I see my connection statistics?”. If you need more, you can purchase additional connections. To learn what connections are and how to estimate your usage, see “What are concurrent connections?” and this article.

Yes, if your plan's limit isn't enough. You can buy them automatically in your Dashboard — no support ticket needed.

When you can add them:
— at checkout — add connections in the configurator along with your IPs and other options;
— while your service is active — if you find you need more, add them from your active service page.

Connections are sold in blocks of 1,000. Payment comes from your balance, and the limit increases as soon as the payment goes through. When you add connections to an active service, the charge is prorated for the days remaining in your current billing period, so you pay only for the rest of the term. You see the exact price before you confirm.

Before you buy, check two things. First, look at your connection statistics to confirm you regularly hit the limit. Second, make sure your software closes connections when it's done. Sometimes stuck connections take up the limit, and once the software is configured correctly, you don't need to buy more (see “What are concurrent connections?”).

If several IPs are whitelisted for a service (say, a server, an office, and a home computer), they share one concurrent connection limit by default. A single busy device can then take up almost the whole limit and leave the others short.

To prevent this, set a quota for each IP:
1. Open your active service in the Dashboard.
2. Go to “Management” → “IP Whitelist.”
3. Set how many connections each IP can use.

For example, with a 1,500-connection limit, you can allocate 1,000 to the server, 300 to the office, and 200 to the home computer. The server then can't use up the office's connections, and vice versa.

You can change quotas at any time and as often as you like — for example, give the server more connections temporarily for a large job. Your service's total limit stays the same. If it isn't enough, you can buy additional connections.

Open your active service in the Dashboard, go to the “Statistics” tab, and select the “Connections” chart.

The chart shows how many concurrent connections your service used over time and what its limit is. It helps you see:
— whether your plan's limit is enough or you hit it regularly;
— what time of day your load peaks;
— whether your software keeps connections open when no real work is happening — for example, if the chart stays flat after your tasks have finished.

If the chart often reaches the limit, you have three options: optimize your software (see “What are concurrent connections?”), adjust the quotas across your IPs, or buy additional connections.

Your service (not your entire account) will enter “Slow Mode.” In this mode, you're allowed no more than one concurrent connection per IP address in your service package. The mode is dynamic:

Up to 2 violations per day: 5 minutes.
From 3 to 6 violations: 15 minutes.
From 7 to 10 violations: 30 minutes.
More than 10 violations: 60 minutes in “Slow Mode.”

We explain more about threads and connections in this article.

Our rotating proxies are designed to work with third-party APIs. One API request = 1 credit from your plan. We count all requests based on concurrent connections. If you're working with standard websites (e.g., page scraping), one request (thread) isn't always the same as one connection. Your browser or program may send dozens of requests at once. For example:

- Loading an online store page may take 10 to 40 requests.
- A Google search can send up to 100 requests per query.

Why does this happen?
Every webpage consists of multiple elements: text, images, styles, scripts, and so on. Each element is loaded through a separate request. So opening a single page might generate 10, 20, or even 50 requests — these are concurrent connections. It all depends on the specific site and your software, which is why we can't tell you in advance how many concurrent connections will be used per thread.

When choosing a plan, keep in mind that one request to a site can generate dozens of separate connections. This matters for correctly estimating your request volume and limits.

How do I check how many requests a site makes?
Before choosing a plan, we recommend checking how many requests the target website makes. You can do this with online tools such as RequestMap: enter the URL and you'll get a map of all requests, with their number and types. For example, it shows that the page https://papaproxy.net/faq.php consumes 21 requests (credits) when opened in a browser.

No, we only break proxies down by country. There's no way to select a specific city.

You can't pick specific subnets manually — the list is generated automatically, but from a large pool: we already have over 400 class C subnets (in the 146.19.39.0/24 format), so you get the widest possible variety within a package.

The rough number of /24 subnets depends on the package size and how popular the geo is (the more popular the country, the more subnets):
— a 100-IP package — from 3 to 30 subnets;
— a 500-IP package — from 5 to 40 subnets;
— a 1,000-IP package — from 50 to 150 subnets;
— a 5,000-IP multi-country mix — from 150 to 200 subnets.

Yes. Once every 8 days, you can regenerate your IP list from your Dashboard to get new addresses. This is optional, but available whenever you want it. We can't predict what percentage of the IPs will be new, since the process is fully automated.

This option isn't available for rotating proxies: you don't see the IP list and work through our entire proxy pool. For dedicated proxies, an IP refresh is available once every 30 days.

If you don't want to wait 8 days, you can replace IPs sooner for a fee — see “Can I replace IPs before the next free refresh?”.

Yes. A free proxy list refresh is available once every 8 days (see “I bought a proxy package, but I want to refresh the IP list”). If you need new IPs sooner — for example, a target website has blocked some of them — you can request an off-cycle refresh.

Price — $10 per refresh, deducted from your balance.

How to do it:
1. Open your active service in the Dashboard.
2. Click “More actions” in the top-right corner and select “Refresh IPs.”
3. Choose the refresh type:
— full — the entire IP list for the service is replaced;
— partial — only selected IPs or the IPs from a specific country are replaced, and the rest stay as they are.

A partial refresh is useful when most IPs work well and only a few are causing problems: the working IPs stay in place, so you don't need to reconfigure your software for them.

Within the first 24 hours after purchase, we replace IPs or subnets that aren't a fit at no charge — up to two replacements per service (see “If I buy two identical packages…”). Rotating proxies don't need replacement, since the exit IP rotates automatically.

No, the lists won't be identical: each package is generated randomly from the available addresses in our pool, and the system aims for the widest possible IP diversity. That said, some overlap between packages with the same location is possible — it depends on the size of our pool in that country and current network load, so we can't guarantee there will be no shared IPs. If any IPs or subnets aren't a fit for your use case, we'll replace them free of charge within the first 24 hours after purchase — up to two replacements per package.

If your use case calls for fully unique packages with no overlapping IPs at all, we can build them manually to your exact requirements. This is a paid add-on service — see our Terms of Use for details.

To check the geolocation of your datacenter proxies, use MaxMind (GeoIP) — the industry-standard database that most websites, anti-fraud systems, and IP-lookup tools rely on. MaxMind is our default geo source: packages are assembled by countries as MaxMind records them.

Keep in mind that datacenter proxies can change their geolocation records more often than residential or mobile proxies, and other databases — RIPE, 2IP, IP2Location, DB-IP — update on their own schedules, so for some addresses they may temporarily disagree with MaxMind.

We rely on MaxMind data as the default source, as stated in our Terms of Use.

Note that major tech companies — Meta, Google, Amazon, Microsoft, and Apple — maintain their own geolocation databases and don't rely on third-party sources. If you need a package verified against a specific database — RIPE, IP2Location, or another source — open a ticket in your Dashboard. This is a paid service that lets us tailor a package to your specific geolocation requirements.

Check things in order of priority:

1. IP whitelisting. Make sure the proxy list is whitelisted to the current IP of the device you're working from. If your main IP has changed, re-whitelist your new IP in your Dashboard. You can't verify this through a proxy checker: the proxies are only reachable from the whitelisted IP.
2. Connection limit. Check whether you've exceeded your plan's concurrent connection limit (shown in your Dashboard). If you have, the service enters Slow Mode — see “What happens if I exceed the limits?”.
3. A block on a specific site. If whitelisting and limits are fine, check whether the IP is blocked on your target.

The quickest way to see the error code is a simple request from the terminal on the device the service is whitelisted to. Test the proxy itself, then a request through the proxy to a specific site:

curl -x http://USERNAME:PASSWORD@PROXY_IP:8080 https://ipinfo.io/ip
curl -x http://USERNAME:PASSWORD@PROXY_IP:8080 -I https://target-site.com

If you see the proxy IP and code 200, the proxies work and the issue is with your software. Code 407 — an authentication or whitelisting issue; 429 — the connection limit is exceeded; 403 — a block on the site's side. If the problem is on our end, send us the command output through support.

Your proxy list works only from the IP address you've whitelisted in your Dashboard. Since a proxy checker runs from a different IP than yours, the proxies will appear unavailable to it.

You should only test proxies from the device whose IP is whitelisted to the package in your Dashboard. From any other IP — through online checkers or third-party tools — the proxies won't work: that's by design, not a malfunction.

Step by step:
1. Whitelist the package to the real IP of the device you'll test from (disable any VPN first). To find your IP and set up whitelisting, see “What is IP whitelisting and why is it needed?”.
2. Enter the proxy IP and port in your browser settings (for HTTP we recommend port 8080). Step-by-step guides: setup in Google Chrome and setup in Mozilla Firefox.
3. Open whoer.net (or ipinfo.io). If the site shows the proxy IP instead of your real address, the proxies are working correctly.

For a quick check without a browser, run a curl request from the same whitelisted device — the commands are in “What should I do if my proxies stop working?”.

Without exception, all of our proxies are anonymous — your real IP is fully hidden. If it's still visible during a check, the cause is almost always a setting on your end.

The most common culprit is a WebRTC leak in the browser: even when traffic goes through the proxy, WebRTC can expose your real IP directly. Disable WebRTC (via an extension or in your browser/antidetect settings) and check again. Also make sure the proxy is applied in the right application, not just system-wide, and that any VPN is turned off. Less often, the cause is a faulty checker.

We block mail ports (SMTP, POP3, IMAP) to prevent email spam through our proxy servers. Spam and any illegal activity through our proxy servers are prohibited, as stated in our Terms of Use.

We've also blocked most of the popular proxy checkers. Your proxy package is whitelisted to the IP address of the device (computer or server) you work from, so a request from the proxy checker's external IP doesn't go through — by default the proxies are unavailable to it. In addition, we block Google to avoid complaints from the search engine — please don't use our proxies with this target. You can find a more detailed list of what we block in our Terms of Use.

Everything you do manually in the Dashboard can also be done in two other ways: through the API and through MCP. Both are included in every account at no extra cost.

The API lets you manage your proxies from your own code — a script, a bot, or an internal service. Instead of clicking through the Dashboard, your software sends the requests. For example, it can check which countries have IPs available, calculate a quote, buy or renew a service, update IP whitelisting, download your proxy list, or pull usage statistics. It's useful when tasks repeat regularly or need to be part of your own workflows.

MCP (Model Context Protocol) is a standard that AI agents such as Cursor and Claude Code use to connect to external services. Once you connect our MCP server, you describe what you need in plain English — for example, “How much would 1,000 ISP proxies in the US cost for 3 months? Don't buy anything yet” — and the agent takes the necessary steps. No coding required.

How to create a key:
1. In your Dashboard, open “Settings” → “API Keys”.
2. Give the key a clear name (for example, “Scraper” or “Cursor”) and select only the scopes it needs — see “Which API key scopes are available?”.
3. Copy the key right away: the full key is shown only once.

Security. Store the key in a password manager, an environment variable, or your MCP client's configuration. Don't publish it in code, a repository, or chat messages. Create a separate key for each integration so you can revoke one without affecting the others.

Request examples, Cursor and Claude Code setup, and the full feature list are on our API and MCP for developers page.

Scopes define what a key can do. A request succeeds only if the key has the required scope. Scopes apply the same way to the API and to an AI agent using MCP — for example, a key without wallet:spend can't pay for anything through either interface.

The main rule: give each key only what it needs. Typical setups:
— monitoring and analytics: services:read and usage:read. This key can see everything about your services but can't change or spend anything;
— provisioning: catalog:read, services:read, services:write, wallet:read, and wallet:spend. This setup is for scripts and agents that order and configure proxies on their own.

About wallet:spend. This is a financial scope: software or an agent with it can spend your account balance. Your card details stay with the payment processor — payments come only from your account balance. Keep the balance at a level you're comfortable letting automation spend. For each key, you can see when and from which IP it was used, and you can revoke a key at any time in your Dashboard.

Learn more in the Scopes section of our developers page.

A webhook is an automatic notification we send to a URL you specify whenever something happens in your account. For example, a payment goes through, your balance is topped up, a new service is provisioned, a service is about to expire, or your balance is too low to cover a renewal.

Here's the easiest way to think about it. Without webhooks, to find out that a service is about to expire, you'd have to log in to the Dashboard or write a script that asks the API “anything new?” every few minutes. A webhook works the other way around: as soon as the event happens, we send a message to your URL. All you have to do is react to it.

Common uses:
— accounting — record payments and top-ups in your CRM or bookkeeping system automatically;
— automation — update the proxy list in your software as soon as a service is provisioned or renewed;
— alerts — forward notifications to Telegram, Slack, or email so your team stays informed without opening the Dashboard;
— balance control — top up in time so your services aren't interrupted.

How to set it up:
1. In your Dashboard, open “Settings” → “Webhooks.”
2. Enter the HTTPS URL where your service accepts requests and choose the events you want.
3. Save the secret — it's shown only once. Your service uses it to verify that a notification really came from us and not from a third party.
4. Send a test event and make sure your handler received it.

The Dashboard shows the delivery history for each URL: what was sent and how your server responded. If your handler fails, you can redeliver a specific event.

Don't have your own server? You can receive webhooks through no-code tools such as Zapier, Make, or n8n and forward them from there — for example, to Telegram or Google Sheets. Managing webhooks requires a key with the webhooks scope — see “Which API key scopes are available?”. Technical details are on our developers page.