NetNut alternative in 2026: what to use after the seizure

NetNut sold proxy bandwidth. Residential, static residential, mobile and datacenter IPs, billed per gigabyte, aimed at teams that already ran their own scrapers and browsers and only needed better exit IPs. Spider Cloud is a research lab for web data quality that ships a scraping API, a hosted browser, and a proxy mode you can point any HTTP client at. The two overlapped on exactly one thing, the proxy, and that overlap is the reason people land on this page.

The verdict changed on July 2, 2026. Krebs on Security reported that the FBI seized hundreds of NetNut domains that day, and on September 2 the pricing and product pages still return a seizure banner. If you were routing production traffic through NetNut, the question is no longer which proxy is cheaper per gigabyte. It is what to plug in this week so the crawls keep running, and whether to replace only the proxy or the stack you built around it. Spider covers both cases: the proxy mode is a drop-in host swap, and the scraping API removes the browser fleet the proxy was feeding.

What happened to NetNut

The facts below come from two sources, Krebs on Security and a press release from NetNut’s parent company. We have not added to them.

On July 2, 2026, the FBI seized hundreds of NetNut domains. Krebs on Security reported that the exit pool was alleged to overlap with the Popa botnet, a network of roughly two million devices, and that Alarum Technologies, NetNut’s Nasdaq-listed parent, disputes that account.

On July 4, Alarum announced it would “temporarily pause traffic through the relevant network services for several days” and warned that “availability of the Company’s services will be significantly reduced.” As of September 2, netnut.io, its pricing page and its rotating residential page all show the seizure banner. We could not find a vendor statement with a restart date.

For anyone on the customer side, the practical outcome is the same regardless of how the dispute resolves. The gateway you configured months ago is not answering, and the vendor cannot tell you when it will.

Side by side

NetNut’s row reflects the last state we could document. The vendor’s own pricing page has been unreachable since July, so every NetNut price on this page is a third-party listing and we say so each time.

Pricing

NetNut, as listed by third parties

There is no current NetNut price list. What exists are review-site snapshots, and they disagree with each other.

Proxyway’s review, dated April 2025, listed rotating residential at $99 for 28 GB, which works out to $3.53 per GB, ISP proxies at $99 for 7 GB, mobile at $99 for 13 GB and datacenter at $100 for 100 GB. ProxyLook’s listing, checked on September 2, 2026, shows a Starter tier at $3.45 per GB with a 28 GB minimum, a Pro tier at $3.00 per GB with a 100 GB minimum, and notes there was no true free trial, only a seven day refund window after a sales call. Older listings on other review sites quote higher per GB figures with larger minimums. None of these pages are the vendor, and none of the plans can be purchased today.

The historical pattern is clear enough even from third parties. NetNut wanted a monthly commitment of about $99 or more, sold bandwidth rather than requests, and did not offer pay as you go. As of September 2026 there is no vendor pricing page to link.

Spider

Spider bills two things. Bandwidth at $1 per GB of content transferred, measured before any transformation, and compute at $0.001 per minute of server-side processing. Failed requests cost nothing. Credits are pay as you go, start at $1, and never expire. The pricing and plans guide puts a mixed workload at about $8 to $15 per 100,000 pages in smart mode, from $1/GB plus $0.001 per CPU minute; failed requests cost nothing.

The worked example that matters for a NetNut migration is the proxy mode itself. The same guide lists proxy mode at roughly $1 to $4 per GB depending on pool, on the same account and the same credit balance as the API. If you keep your existing scraper and only swap the proxy host, that is the line item. If you move the scraping onto Spider’s API, bandwidth is measured on the bytes Spider fetched, and the browser fleet you ran beside NetNut goes away.

For teams that want a fixed bill, the Unlimited plan starts at $40 per month for two units of concurrency and scales per unit. Unlimited routes through a smaller flat-rate proxy pool, and geo-targeting is not available on it yet, so a NetNut replacement that needs country pinning should start on pay as you go.

What we could not verify

NetNut never published a benchmark methodology, and a proxy network is not the kind of product that gets one. Success rate through a proxy depends on the client behind it, so any number we quoted would measure our scraper, not their IPs. We are also not going to speculate about the Popa botnet allegation or the legal proceedings beyond what the two sources above say.

On our side, the closest public measurement is the Spider Browser dataset: a 100% pass rate on 999 URLs across 254 domains (February 2026, dataset at https://github.com/spider-rs/spider-browser-dataset). That is our own benchmark, not independently verified, and it measures the hosted browser plus our infrastructure end to end rather than the proxy pool in isolation. The URLs, per-URL results and timings are in the repo so you can rerun it against whatever you use today.

Where NetNut wins

It is strange to write this section about a service that is currently offline, but the comparison is only honest if we say what NetNut did better while it ran.

NetNut sold the proxy and nothing else, and for some workloads that is the right shape. Ad verification, brand monitoring, geo-testing a checkout flow, or any job where you need the raw upstream response with no rendering or transformation in the path, wants a plain pipe. Spider’s proxy mode can do that, with tracker blocking on by default and a flag to turn it off, but the product is built around returning clean data, and a team that only ever wanted bytes will notice the extra features it is not using.

NetNut also offered mobile and datacenter pools alongside residential and ISP. Spider’s proxy mode has residential and ISP. If a workload specifically needs mobile carrier IPs, Spider is not the like-for-like replacement for that slice.

Finally, per GB pricing with large commitments can be cheaper than usage pricing at very high, very steady volumes, once the sales call is done. Third-party listings put NetNut’s committed rates near $1 to $2 per GB. Spider’s proxy mode starts at $1 per GB with no commitment, so the gap at the top end was small, but it existed.

Where Spider wins

The obvious one first: Spider’s proxy mode is answering requests today, on the same host and port it was on last quarter, with no sales call required to start at $1.

Beyond availability, the difference is what happens after the proxy. A proxy changes the IP a request arrives from and nothing else. Rendering, challenge pages, retries, and turning HTML into the format your pipeline wants stay with you. Teams on NetNut typically ran a browser fleet, a retry layer and a parser to cover that, and each of those had its own bill and its own on-call. Spider’s API returns markdown, text or structured JSON from a single request in http, smart or chrome mode, and the unblocker endpoint handles sites behind common bot-protection services without per-site configuration. The proxy is still there when you want it, but most migrations stop needing it.

Cost follows from that. Failed requests cost $0, so a blocked page does not consume the bandwidth a proxy would have billed for. Credits never expire, so a team that crawls in bursts is not paying a monthly minimum for idle months. And the 10,000 requests per minute default limit means a catalog crawl that took an evening through a proxy-and-browser setup fits in a lunch break, subject to how fast the target site will serve.

The proxy mode itself is also more than a host swap. Options travel in the password field, so proxy=residential or proxy=isp picks the pool per request, country_code=de pins the exit country, and trackers_disabled=false turns off the default tracker blocking when you need every byte. Any client that supports a proxy works, and the same API key runs the REST endpoints, so you are not maintaining two vendor accounts.

Migration: the same request through both

The task is the same in both blocks: fetch a product page through a residential exit in Germany, from Python with requests.

Through NetNut, as it was configured

import os
import requests

# NetNut gateway from an older integration; the host no longer answers.
proxies = {
    "http": f"http://{os.getenv('NETNUT_USER')}-cc-de:{os.getenv('NETNUT_PASS')}@gw.netnut.io:5959",
    "https": f"http://{os.getenv('NETNUT_USER')}-cc-de:{os.getenv('NETNUT_PASS')}@gw.netnut.io:5959",
}

response = requests.get("https://example.com/products/123", proxies=proxies, timeout=30)
html = response.text
# Render, retry, and parse are still your job from here.

Through Spider’s proxy mode

import os
import requests

key = os.getenv("SPIDER_API_KEY")
options = "proxy=residential&country_code=de"

proxies = {
    "http": f"http://{key}:{options}@proxy.spider.cloud:8888",
    "https": f"https://{key}:{options}@proxy.spider.cloud:8889",
}

response = requests.get("https://example.com/products/123", proxies=proxies, timeout=30)
html = response.text

Two lines change: the host and the credential format. The username is your API key and the password carries the options. Full parameter list in the proxy API reference.

Or skip the proxy and ask for the data

If the reason you had a proxy was to feed a headless browser, this is the version that removes the browser too.

import os
from spider import Spider

client = Spider(api_key=os.getenv("SPIDER_API_KEY"))

pages = client.scrape_url(
    "https://example.com/products/123",
    params={
        "request": "smart",
        "return_format": "markdown",
        "country_code": "de",
    },
)

print(pages[0]["content"])

request: "smart" starts with a plain HTTP fetch and moves to a rendered fetch only when the page needs it, which is where the $8 to $15 per 100,000 pages figure comes from. Set request: "chrome" to force rendering, or route the request through the unblocker endpoint for sites that block the standard path.

When to pick which

Pick a plain residential proxy provider, Spider’s proxy mode included, when you already have a scraper you trust and the only thing that broke in July was the exit IP. The migration is a host swap and you keep everything else.

Pick Spider’s API when the proxy was one of three or four services holding up a scraping pipeline and you would rather not run the other three. Start in smart mode, check the returned markdown against what your parser produced, and turn on chrome mode or the unblocker only for the domains that need it.

Wait, and keep a fallback in place, if your workload needs mobile carrier IPs or a datacenter pool at the bottom of the price curve. Spider does not sell those pools, and no one can say today whether NetNut will.

For the wider field, the compare page lists every vendor we have written up with the same table as above, and the posts on Bright Data and Oxylabs cover the two proxy-first vendors most NetNut customers are evaluating alongside this one.

Whichever way you go, run the same URL through the playground first. It shows the returned content, the mode that served it and the cost of that one request, which is the fastest way to know whether smart mode covers your targets before you move a single crawl.