Shifter vs FleetProxy
Read this first: FleetProxy is not a separate network from ProxyEmpire. Both hostnames resolve to the same nine gateway IP addresses on the same port, use the same username format, and on runs taken the same day they served 57% of the same exit addresses, rising to 90% in Germany. With that said, we sent the same requests through both networks in five countries. Shifter returned more live residential IPs in every market, answered roughly twice as fast, and costs $0.75 per GB against $2.12.
Trusted by 50,000+ clients worldwide
Every provider in this series gets the same test: plain HTTP, one fresh connection per request, 1,000 concurrent, 10 second timeout. FleetProxy completed between 49.5% and 67.6% of requests depending on the market. Almost all of the shortfall was redirects, between 21,000 and 85,000 per run, with connection resets making up the rest. The counts below are from those runs exactly as measured. We did not switch transport or lower concurrency to lift the number, because a client that loses a third of its requests is something you would rather learn here than in production.
Same test, a third of the price
Compared at 100 GB a month, a volume a working team actually buys: FleetProxy is $2.12 per GB there and Shifter is $1.00, falling to $0.75 from 200 GB. FleetProxy's $1.59 floor needs a 500 GB purchase at $795.
| Monthly volume | Shifter | FleetProxy | You save |
|---|---|---|---|
| Entry | $10 5 GB at $2.00/GB | $2.65/GB 1 GB and up | 25% |
| 100 GB | $100 $1.00/GB | $212 $2.12/GB | 53% |
| Best rate | $0.75/GB from 200 GB | $1.59/GB from 500 GB, $795 | 53% |
Live IPs online during the test
Distinct IPs each network had live and reachable at test time, not the size of its total network. FleetProxy's figures come from runs where a third to half of requests failed, so its true pool is larger than shown. Residential IPs rotate all day, so both providers serve far more addresses across 24 hours.
Every metric, country by country
Pick a country to see the complete measurement, including the numbers that did not go our way.
| Metric | Shifter | FleetProxy |
|---|---|---|
| Requests sent | 250,000 | 250,000 |
| Live IPs during the test | 136,667 | 112,402 |
| Unique networks (ASNs) | 1,606 | 1,482 |
| Unique cities | 5,813 | 5,650 |
| Average latency | 364 ms | 922 ms |
| Success rate | 99.8% | 65.8% |
The closest market of the five. Shifter returned 22% more live US IPs, 136,667 against 112,402, across 1,606 networks against 1,482, and answered in 364 ms against 922 ms. FleetProxy completed 65.8% of its requests, so its real pool is larger than shown here. On price the gap is wider than the pool: $0.75 per GB against $2.12.
How we measured this
Before any of the numbers: FleetProxy and ProxyEmpire share infrastructure. A DNS lookup on residential-v2.fleetproxy.io and v2.proxyempire.io returns the same nine addresses, both answer on port 5000, both issue usernames as r_ followed by ten hex characters, and both show the same intermittent blank responses. Comparing their pools against each other is comparing two front ends onto substantially one network.
Every provider in this series receives the identical test: plain HTTP to an IP echo service, one fresh connection per request, 1,000 concurrent, 10 second timeout, and fixed volumes of 250,000 requests in the United States, 100,000 in the United Kingdom, 80,000 in France, 70,000 in Canada and 60,000 in Germany. One connection per request means one fresh exit per request, so distinct addresses map directly to the reachable pool.
FleetProxy completed 49.5% to 67.6% of those requests. The failures were overwhelmingly redirects, between 21,000 and 85,000 per run, with connection resets making up the rest. We report the shortfall as observed and do not attribute it to a concurrency ceiling or any other single cause, because these runs were not designed to isolate one.
Both providers ran plain HTTP, so latency is directly comparable on this page. Because FleetProxy's counts come from partial runs, its true live pool is larger than the figures shown. Its own gateway also shares hosts and exit addresses with ProxyEmpire, which is covered below.
FleetProxy and ProxyEmpire resolve to the same set of gateway hosts on the same port, use the same username format, and on runs taken the same day returned 57% of the same exit addresses, rising to 90% in Germany. If you are comparing the two against each other, you are comparing two front ends to substantially one network.
What these numbers mean
Every count on this page is the number of IPs that were live and reachable at the time of the test, not the size of either provider's network. Residential IPs rotate constantly as real devices come online and drop off, so across a full day both networks serve far more addresses than a single window can capture. What limits a crawl is how many respond at the same moment, which is what this measures.
Why use Shifter residential proxies?
A larger live pool in all five markets
22% more in the United States, 17% more in the United Kingdom, 73% more in Canada, 79% more in Germany, and nine times the pool in France.
Roughly twice as fast
364 ms against 922 ms in the United States, 324 ms against 994 ms in Canada. Both ran plain HTTP, so these timings compare directly.
Completes the test
FleetProxy answered 49.5% to 67.6% of requests. Between a third and half came back as redirects or resets on the standard plain HTTP test.
A third of the price
$0.75 per GB from 200 GB against $2.12 at 100 GB, or $1.59 if you commit to 500 GB at $795. Every Shifter tier is published without a sales call.
Frequently asked questions
Common questions about how we benchmarked FleetProxy.
We can only report what we measured, and what we measured is this: the two resolve to an identical set of gateway hosts on the same port, use the same username format, and on runs taken the same day they served 57% of the same exit addresses, rising to 90% in Germany. Two genuinely separate networks we tested the same way shared 0.1%. Draw your own conclusion, but do not treat them as two independent options.
Almost all of the failures were redirects, between 21,000 and 85,000 per run, with connection resets making up the rest. Single requests succeed reliably, so the shortfall appears only under concurrency. We are reporting what we measured and not guessing further, since these runs cannot separate a capacity limit from a configuration one.
It makes it identical, which is the point. Every provider in this series gets the same test, and any of them that cannot complete it is shown not completing it. Because FleetProxy's runs were partial, its true pool is larger than the figures here, which we say plainly on the page.
The United Kingdom, where it edged Shifter on network spread at 194 to 193 and covered more cities, 1,971 against 1,842. Its geo accuracy was also excellent everywhere, 99.5% to 100%, better than ours in the United States. What it does not match is pool depth outside Britain, response time, or price.
A third of the price, a deeper pool
More live addresses in every market we measured, answering roughly twice as fast, at $0.75 per GB against $2.12. Run the same test yourself and compare.