There are two ways to pay for proxies: per IP address for a term, or per gigabyte of traffic. With the first you buy a list of addresses for a month and use it however you like. With the second you buy a traffic package and create the proxies yourself — as many as you need, in the countries you need. Which one is cheaper depends not on the unit price but on how many addresses you need and how much traffic goes through each of them.
How paying for traffic works
Here is how Pay-per-GB works at Netrun:
- You buy a package of gigabytes, and it becomes your traffic pool. Packages you buy later are added to the same pool.
- In the Pay-per-GB panel you create proxies: pick a country and a quantity — up to 500 at a time. The panel is in your account on the website ("NETRUN proxies" → traffic) and in the Telegram bot.
- You download the list in the protocol (SOCKS5 or HTTPS) and line format you need and use it like any other proxies.
- The pool is charged for the traffic that actually went through the proxies, in both directions: your requests and the websites' responses. The remaining balance and the pool's term are shown in the panel.
- Need fresh addresses or another country? Run a new generation.
Pay-per-GB proxies are IPv6 proxies, so they work with websites that have IPv6. Checking a website takes a minute — see "Which Websites Support IPv6".
How the two models differ
| Per-IP IPv6 | Pay-per-GB | |
|---|---|---|
| What you pay for | Each address for the plan's term (30–180 days) | Gigabytes of traffic |
| How much traffic | Doesn't affect the price | Exactly what you use |
| How many addresses | As many as you bought | As many as you create, up to 500 per generation |
| IP addresses | Assigned to you for the whole term | New ones with each new generation |
| Country | Chosen when you buy | Chosen at every generation |
| Best for | Accounts that need their own permanent IP | Scraping and monitoring on IPv6 websites |
When paying for traffic is cheaper
Paying per gigabyte wins when you need many addresses but little traffic goes through each one:
- Price and stock monitoring. Checking hundreds of product pages once a day or once an hour means many requests, but each one is light.
- Checking search results and rankings in different countries.
- Collecting data through APIs, where a response is a few kilobytes of JSON rather than a heavy page full of images.
- One-off and irregular jobs. No need to keep a list of addresses for a month for a couple of days' work.
- Different countries for different jobs. You choose the country at generation instead of buying a separate list for each one.
When per-IP proxies are cheaper
- Accounts and multi-accounting. An account needs the same IP day after day, and a per-IP address stays assigned to you for the whole term.
- Lots of traffic per address. Video, downloads, browser profiles that run all day — paying per gigabyte for that volume would cost more.
- A predictable budget. You know the price in advance, and it doesn't depend on how much data goes through.
How to estimate which one fits your task
Work out two totals and compare them.
- Per IP: the number of addresses you need at the same time × the price of one address per month.
- Per traffic: requests per month × the size of one response → gigabytes × the price per gigabyte.
To estimate response sizes:
- a full page load in a browser, with images and scripts, is usually 1–5 MB;
- the HTML of a page without images is tens to hundreds of kilobytes;
- a JSON API response is usually a few to a few dozen kilobytes.
Example: you check 3,000 products every day and download only the HTML, about 300 KB each. Over a month that is 3,000 × 30 × 0.3 MB ≈ 27 GB. If an API returns the same data at 20 KB per response, it comes to under 2 GB. The same task can cost several times more or less depending on what exactly you download, so estimate the response size first and then pick the pricing model.
The most reliable way is to buy a small package, run the task for a day or two and look at the usage in the panel. Scaling that up to a month is easy.
How to use less traffic
- Turn off images, video and fonts in your scraper or browser if you don't need them.
- Take data from an API or from the JSON embedded in a page rather than the full page.
- Allow compressed responses with the
Accept-Encoding: gzip, brheader — text shrinks several times over. - Don't download the same thing twice — cache whatever doesn't change.
FAQ
What happens when the traffic runs out?
The proxies stop passing traffic until you buy another package. The new package is added to the same pool.
Can I run accounts on Pay-per-GB?
You can, but it isn't the best choice: accounts need a stable address and usually a fair amount of traffic per session. Per-IP IPv6 proxies suit them better — the address stays yours for the whole term.
Is used traffic refunded?
No: you are charged for what actually went through the proxies, and that can't be returned. So first test on a small volume that the website opens and the task works as it should.
You can buy a traffic package on the home page under the "Traffic · GB" tab, and a list of addresses for a term in the IPv6 configurator.