Atlas disclosure
Shifter runs Atlas. How it stays fair.
We sell proxies and we rank proxy networks, Shifter included. This page says how we keep the two apart.
Shifter
Runs Atlas, and is ranked in it
Shifter sells residential and ISP proxies and runs Atlas. Shifter is tested with the same requests, targets and formula as every other network, and its results are published the same way, including when they are worse.
- Same requests
- Same targets
- Same formula
- Same tests for Shifter
- No paid placement
- Public corrections
- Open data
Ranked by Atlas Score across 18 networks
Shifter94.1#1
Oxylabs83.4#2
Decodo81.5#3
FleetProxy80.6#4
Geonode80.3#5
DataImpulse80.1#6
ProxyEmpire79.9#7
Webshare79.3#8
NodeMaven77.7#9
SOAX77.5#10
Proxy-Seller77.2#11
Byteful75.6#12
Infatica74.6#13
Massive74.2#14
Evomi73.1#15
Rayobyte71.5#16
Proxy-Cheap70.5#17
IPRoyal66.4#18
Shifter94.1#1
Oxylabs83.4#2
Decodo81.5#3
FleetProxy80.6#4
Geonode80.3#5
DataImpulse80.1#6
ProxyEmpire79.9#7
Webshare79.3#8
NodeMaven77.7#9
SOAX77.5#10
Proxy-Seller77.2#11
Byteful75.6#12
Infatica74.6#13
Massive74.2#14
Evomi73.1#15
Rayobyte71.5#16
Proxy-Cheap70.5#17
IPRoyal66.4#18The rules we hold ourselves to
Six commitments that keep Atlas fair when the company running it is also in it.
-
Same tests for Shifter
Shifter gets the same requests, targets, schedule and formula as every other network. There is no separate setup and no adjustment.
-
No paid placement
Position in any table comes from the published formula alone. There is no field in the dataset that a payment could change.
-
We buy access like a customer
Networks are tested through access bought the way any customer buys it, not through a plan set up for the test.
-
A right of reply, not of edit
Any network can dispute a figure. We measure again and publish the outcome either way, including when we were wrong.
-
Corrections in public
If a published number changes, the change and its cause are logged with a date. Nothing is changed quietly.
-
Open data
The dataset behind every page downloads as one file, so any figure can be checked against the formula.
The formula behind every rank is on the Methodology page, and the ranks it produces are on Atlas Overview.
Next · 02 Open source Can anyone get the same results?Every test is open source
You do not have to trust our numbers. The code that measures every network, Shifter included, is public, so anyone can run the same tests and get the same results.
Atlas is only as fair as it is checkable. So the whole method is in the open.
- The collector that sends every request
- Every target and its written pass rule
- The formula that turns results into a rank
- The dataset every page is built from
Replicate it in three steps
- 1
Get the framework
The same code that runs Atlas, for every network, Shifter included.
git clone https://github.com/shifter-io/atlas - 2
Run it on your own account
Run each target of `atlas targets` once with the provider's config, as the README shows. Your credentials stay on your machine.
- 3
Compare with ours
Put your figures next to the published dataset. If the shapes match and only the absolute numbers differ, the gap is your location, not the method.
Who is behind Atlas
Every party with a stake in Atlas, and how it relates to the rankings.
Testing without doing harm
Measuring proxies means sending automated traffic to other people's sites. These rules limit how we do it, which is one reason some numbers are deliberately conservative.
-
Low request rates
Request rates are deliberately low, rate-limited per target, and respect robots directives where they apply to automated access.
-
Real people's connections
Residential exits belong to real people. A measurement must not become a burden on someone else's connection.
-
Read-only, public pages
Targets represent common commercial use, and each test is a read-only fetch. We never request authenticated sessions, submit forms or bypass a paywall.
-
A block is a finding
A challenge or a block is recorded as it is. We do not solve CAPTCHAs, rotate around a ban, or try to defeat a target's defences.
-
No exit IPs published
We never publish full exit IP addresses, on the page or in any download. The public record is counts, ASNs and location.
-
Credentials stay put
We do not resell, aggregate or retain provider credentials beyond the machine that runs the collector.
Changes to how we test are logged in What changed.
Next · 05 Questions What do people ask about this?Questions we expect
Straight answers to what people ask when a proxy company ranks proxy networks.
Who runs and funds Atlas?
Shifter runs and pays for Atlas, and is ranked in it like any other network. We sell residential and ISP proxies, and have run residential proxy infrastructure since 2012.
Can Shifter rank itself first?
Only by measuring best. Every network, Shifter included, gets the same requests, targets and formula, and the rank comes from the formula alone. Shifter's results are published the same way when they are worse.
Can a network pay to appear or to rank higher?
No. Rankings come only from the published formula, and there is no field in the dataset that a payment could change.
How can I check the numbers myself?
Two ways. Download the dataset the pages are built from and apply the formula on the Methodology page, and you should get the same ranking. Or run the open-source test framework on your own proxy account and compare your results with ours.
Where do I report an error or dispute a figure?
Send it through the Contribute page. We measure again and publish the outcome either way.
Something else? Ask through the Contribute page.
Same rules for us, so check us
Shifter is measured by the same tests as the networks it competes with. Order with a 3-day money-back guarantee and see whether our published figures hold up.