What “Private Proxy” Means Across Providers (It Varies)

1. Not public. In the early days of proxy lists, "public proxies" were open servers anyone could find and use, often misconfigured machines. A "private proxy" was simply one you paid for and accessed with a password or an allowlisted IP. Many sellers still use the word this way. Under this meaning, a private proxy can be shared by dozens of paying customers. Our comparison of free vs paid proxies covers why the public kind is a bad idea for almost anything.

2. Fewer users than shared. As vendors started selling three sharing tiers (shared, a middle tier, dedicated), some named the middle tier "private" or "semi-private", because it had fewer users than the cheap shared tier. Under this meaning, a private proxy is shared, just not by many.

3. Exclusive to you. The meaning most buyers assume: one IP, one customer, for the length of your subscription. Vendors using this meaning often say "dedicated" and "private" interchangeably on the same page.

All three are in use today, sometimes on competing product pages that rank for the same keyword.

Sources: iproyal.com/private-proxies, help.webshare.io (proxy type article), Rayobyte support portal (semi-dedicated article), myprivateproxy.net, docs.brightdata.com proxy FAQ. All read September 2026.

Three patterns stand out.

The same word, opposite meanings. "Private" is the exclusive tier at IPRoyal and the middle, shared tier at Webshare. A buyer comparing a "private proxy" price across the two is comparing different products.

The middle tier's size varies. Webshare's middle tier is two users. Rayobyte's is up to three. A vendor without a published number could mean anything.

Exclusivity can have a scope. Bright Data's documentation ties dedicated IP exclusivity to the target domains you list, and there is no "all domains" option. That is a sensible design for their network, and it also means "exclusive" answers "exclusive for which sites?" rather than "exclusive, full stop."

A brand name is not a sharing policy. MyPrivateProxy has "private" in its name and also sells a shared tier, stated plainly on its own site. Nothing wrong with that, but it shows why the word on the logo or the product card tells you nothing about the IP you receive.

"Private" is the most confusing label, but not the only one. These turn up next to it on product pages, and each needs the same follow-up question.

• Exclusive. Usually means one customer, but check the scope. As above, exclusivity can be tied to named target sites rather than every destination.
• Private residential or private static residential. The sharing tiers carry over to residential and ISP products. Webshare's own blog describes its private static residential product as IPs "shared with at most two individuals", which is the semi-dedicated meaning applied to a different network type.
• Individual or personal proxies. Common on sellers whose main market is not English-speaking, and generally a translation of "dedicated". Confirm it, since translated product pages are where labels drift most.
• Private pool. Used for rotating products where a pool is reserved for a customer segment or an enterprise account, not for one buyer's exclusive IP list. It says something about who the neighbours are, not that there are none.
• Fresh, virgin or unused. A claim about history, not about sharing. An IP can be fresh and shared, or dedicated and heavily used by its previous owner.

The practical rule is the same for all of them: translate the label into a number of concurrent customers, a scope, and a history before comparing prices.

People search for a "private anonymous proxy" as if it were one feature. It is two independent properties, and a proxy can have either without the other.

• Private (in any of the senses above) describes who else uses the IP.
• Anonymous describes what the proxy reveals in your requests: whether it adds headers such as Via or X-Forwarded-For that expose that a proxy is in use or pass along your real IP. The levels are usually called transparent, anonymous and elite. Proxy anonymity levels explained covers them in detail.

The bottom-left cell is rare with commercial providers but common with self-hosted setups, where a default Squid or Nginx configuration adds forwarding headers. Exclusivity never fixes a header leak, and a clean header setup never makes an IP exclusive. Test the two separately, as shown below.

Even when "private" really does mean one user, it describes a narrow thing: nobody else is sending traffic from that address during your subscription. Several things buyers expect from it are not included.

A clean history. The IP had a life before you. It may have been assigned to another customer last month, or sat in a range that anti-abuse lists flagged years ago. Exclusivity starts at purchase. Reputation does not reset. What is IP reputation explains how that history follows an address.

Exclusive neighbours. Your IP sits in a subnet, often a /24 of 256 addresses, and the rest of that block is probably sold to other customers. Many sites rate-limit or block by range, so a neighbour's behaviour can still reach you. See what are subnet proxies for how range-level blocking works.

A private network type. A dedicated datacenter IP still announces from a hosting provider's ASN. Sites that treat datacenter traffic differently will treat yours the same way, however exclusive it is.

Private traffic. Exclusivity is about other customers. The provider still carries your traffic, so logging and data handling are a separate question for the vendor's privacy policy.

Rotating proxy plans work differently from IP lists. You connect to a gateway, and each request or session exits from an address picked from a large shared pool. Other customers use the same pool by design, so asking whether a rotating plan is "private" does not really make sense. What you buy is access to the pool, and privacy of any single IP is not part of the deal.

For scraping and monitoring that is usually the better product. Spreading requests across many addresses matters more than owning a few. SparkProxy is built this way: datacenter plans sold by concurrent threads (Starter $75/mo for 100 threads up to Plus $440/mo for 1000), with unlimited bandwidth, random per-request rotation or 5-minute auto rotation, and a pool of 1M+ datacenter IPs across 80+ countries. We do not sell those plans as "private proxies", because they are not.

If you need the same exit for a short multi-step flow rather than a permanently owned IP, a sticky session on the gateway (port 11002) holds one exit for a period without making it exclusive to you. The difference between the two is covered in shared vs dedicated datacenter proxies.

Send these to the vendor's sales or support team before a large order, and keep the answers. The replies are more useful than any product page.

1. What is the maximum number of customers who can use one of these IPs at the same time? Ask for a number. "Low" is not an answer.
2. Is that limit per IP or per subnet? Some vendors limit users per IP but sell adjacent addresses in the same /24 to others.
3. Was this IP assigned to another customer before, and is there a cooling-off period between customers?
4. Does exclusivity apply to all destinations, or only to specific target sites?
5. What happens if an IP is blocked or blacklisted on day one: is replacement free, how fast, and how many times per term?
6. Can I choose IPs from different subnets? If you buy 20 private IPs and they all sit in one /24, one range block takes out all 20.
7. Does "private" describe the IP, or my account's credentials? A few sellers use the word only to mean your login is not shared.

If a vendor will not answer questions 1 and 4 in writing, assume the least exclusive meaning.

You cannot see other customers on an IP directly. You can look for traces of them, and you can check the anonymity side precisely.

Test 1: has anyone used this IP recently?

GitHub's REST API allows 60 unauthenticated requests per hour per IP address, and its /rate_limit endpoint reports how many remain without spending one. Run this through the proxy before you send any other GitHub traffic from it:

curl -s -x http://USERNAME:PASSWORD@YOUR_PROXY_IP:PORT \
  https://api.github.com/rate_limit | grep -A4 '"core"'

If remaining is below limit and you have not used GitHub through this IP, someone else has, within the reset window shown in the response. Repeat a few times over a day. A dedicated IP that keeps showing consumed quota is not dedicated in the sense you paid for. The reverse is weaker evidence: a full quota proves only that nobody hit GitHub from that IP in the last hour, not that you are alone on it.

Test 2: what does the proxy reveal?

# Your real public IP, for comparison
curl -s https://httpbin.org/ip

# What a site sees through the proxy
curl -s -x http://USERNAME:PASSWORD@YOUR_PROXY_IP:PORT https://httpbin.org/headers
curl -s -x http://USERNAME:PASSWORD@YOUR_PROXY_IP:PORT https://httpbin.org/ip

Look for Via, X-Forwarded-For, Forwarded or X-Real-Ip in the headers output, and check that your real IP appears nowhere. Note that HTTPS requests pass through a proxy as an encrypted tunnel, so header injection mainly shows up on plain HTTP. Test an http:// URL as well if your workload uses one.

Add a reputation check on day one too. Look up each IP on a couple of public blocklist lookup services and record the results, so you have evidence if you need a replacement.

Pay for one-user IPs when the IP is part of an identity or an agreement:

• Logged-in accounts where a stable, unshared address is part of looking like one consistent user.
• Allowlists. A partner, bank, or client firewall grants access to a specific IP, and you do not want another customer on it.
• Rate limits you have to budget. If a target gives each IP a fixed quota, sharing that IP means sharing your quota.
• Contractual or compliance needs where you must show who controls an address.

Skip it, and buy rotating pool access or a cheaper shared tier, when you are collecting public data across many pages, monitoring prices or search results, or testing localisation. There, the number of addresses you can spread traffic across matters far more than owning any one of them. What is a dedicated datacenter proxy goes deeper on the dedicated side.