Bundled or separate proxy: which is cheaper and which disappears less

The two ways to buy proxy inside an anti-detect browser

Model one: the tool sells you the browser and a monthly traffic allowance in the same plan. Model two: the tool sells you the browser and expects you to bring your own proxy, bought from a specialist provider and configured per profile.

Both models exist across the vendors whose pricing pages we read on 2026-08-12. Multilogin publishes an explicit monthly allowance per tier. GoLogin lists a residential proxy row with no quantity beside it, though the page's own embedded pricing data puts 2 GB on every paid tier without saying over what period. Incogniton advertises built-in proxies without a published quota. Dolphin Anty lists proxy protocol support — HTTP, SOCKS5, SOCKS4, SSH — which is not an allowance at all, it is a statement that the browser can use a proxy you supply. Kameleo and AlterAntiX are bring-your-own by design.

The first practical consequence: you often cannot compare vendors on this axis at all, because most do not publish the number. Where a quantity is not disclosed, treat the bundle as worth zero in your budget until support gives you a figure in writing.

Is a bundled proxy cheaper than a separate one?

It is cheaper in exactly one situation: when your monthly consumption fits inside the allowance. Outside that situation it is not a discount, it is a starter pack, and the moment you exceed it you are buying traffic on the open market anyway — with the added disadvantage that you now have two suppliers instead of one.

So the question is not philosophical. It is arithmetic with one unknown: how many gigabytes a profile burns in a month. Everything else in this article is public pricing.

There is a second question hiding behind the first, and it is the one that decides uncomfortable cases: what happens when you go over. A bundle that stops working when the allowance runs out is a different product from one that keeps working and bills you the overage, and that is a different product again from one that silently degrades. The pricing pages we read do not answer this — it is a support question, and it is worth asking before you commit, because discovering the answer on the last week of a campaign is expensive.

The formula: how much traffic a profile burns per month

Nobody can hand you this number, and anyone who publishes an authoritative "gigabytes per profile" figure without saying what they measured is inventing it. Traffic depends on the platform (an ad manager is heavy, a marketplace back office is heavier, a webmail is light), on session length, on how much media loads and on whether you leave dashboards open with live polling.

What you can do is measure yours, in one week, with three inputs:

• Isolate one profile on a dedicated proxy — one that no other profile touches — so the provider dashboard attributes consumption to it alone.
• Work a normal week. Not a light one and not a stress test. The point is a representative sample, not a best case.
• Read the consumption in your provider dashboard at the end of the week and apply the formula below.

GB per profile per month = (GB consumed in the week / 7) × 30

Then scale it: total monthly GB = profiles × GB per profile per month and monthly traffic cost = total monthly GB × price per GB. Three multiplications and one division. The reason this is worth doing rather than guessing is that the answer routinely lands an order of magnitude away from what people assume — in both directions.

What each bundle actually gives you, per profile

Now apply the same division to the published allowances. This is the table that changes the decision, and as far as we can tell nobody publishes it — for understandable reasons.

A note on the GoLogin rows, because they are the only ones here that did not come off the visible page. The pricing table shows a row reading "Gb resident proxy" with no quantity next to it; the 2 GB figure comes from the page's own embedded pricing data, where the same allowance is attached to every paid tier. The page does not say whether those 2 GB are monthly or a one-off, so treat the per-profile figures as an upper bound and ask support before counting on them.

Read the fourth column twice. The allowance per profile shrinks as the plan grows — from 100 MB down to 34 MB at Multilogin, and from 200 MB down to 2 MB at GoLogin, where the allowance does not grow at all. That is not a criticism of any vendor; it is what happens when the traffic bundle scales more slowly than the profile count, which is the only way to keep a bundled plan affordable. But it tells you exactly what the bundle is for: evaluating a profile, not operating one. Thirty-four megabytes is a login, a look around and a logout.

Where the bundle wins

It genuinely wins in three shapes, and they are not rare.

• Very low consumption per profile. Accounts you touch weekly rather than daily, on light interfaces. If your measured number lands under the implied allowance, the bundle is free traffic and you should take it.
• You do not want a second vendor. One invoice, one support channel, one contract. For a small team without procurement patience, that is worth real money.
• You do not have a provider yet. Buying residential traffic well requires knowing what to buy. A bundle removes that decision at the start, which is a legitimate reason to accept a worse unit price while you learn.

Where the separate proxy wins, and why it disappears less

Beyond price per gigabyte, the structural advantage is that unbundled traffic does not have to evaporate at the end of the month. IPRoyal, for example, states plainly on its pricing page that residential traffic never expires. A monthly allowance inside a browser plan does the opposite: it resets, and whatever you did not spend is gone.

That difference matters far more than it sounds for anyone whose workload is uneven — which is most of this market. A quiet month followed by a launch month, seasonality, a client onboarding, a campaign that runs for three weeks and stops. With a resetting allowance you overpay in the quiet months and overflow in the heavy ones. With non-expiring traffic you buy once at the best tier price and draw down as you go.

• Control over the pool. You choose the provider, the country, whether sessions are sticky, and you can move a single profile to a different provider without changing browsers.
• Better unit prices at volume. The published ladders reward commitment: Decodo goes from US$4.00/GB pay-as-you-go down to US$2.75/GB at 100 GB, the largest tier it publishes. A browser bundle has no equivalent lever.
• Diagnosability. When something breaks, a separate provider gives you its own logs, its own IP tester and its own support, instead of one vendor answering for both the browser and the network.
• No lock-in on the expensive half. Changing browsers is annoying. Changing browsers when the traffic is welded to the browser subscription is worse.

Simulation at 10, 30 and 100 profiles

Prices collected on 2026-08-12, and this is the part most simulations get wrong: there is no single per-gigabyte rate to multiply by. Every provider publishes a ladder, and the rate that applies to you is the one attached to the volume you buy. So each row below is priced at the published tier that matches its volume, with the tier named in the cell. IPRoyal publishes US$7.00/GB at 1 GB, US$5.95/GB at 2 GB and US$5.25/GB at 10 GB, and then nothing until a custom plan at US$1.75/GB that starts at 10 TB — which is why that advertised floor prices none of these scenarios. Decodo publishes US$3.75/GB at 3 GB, US$3.50/GB at 10 GB, US$3.25/GB at 25 GB, US$3.00/GB at 50 GB and US$2.75/GB at 100 GB, its largest published tier. We run two consumption scenarios because the whole point of the formula is that your number decides the answer.

One detail in that table deserves its own sentence, because it is where the money hides. IPRoyal's published ladder stops at 10 GB and does not resume until 10 TB, so every volume between those two points is priced at the 10 GB rate — the 200 GB row is charged at the same US$5.25/GB as the 15 GB row. Volume discounts you cannot reach are not volume discounts. Ask any provider for the rate at your volume in writing, and budget with that number rather than with the one on the landing page.

Now put the bundles next to it. The largest published allowance in our sample is 10 GB per month, on a US$89 plan covering 300 profiles. In the light scenario that bundle covers twenty profiles and not one more. In the working scenario it covers five. Everything beyond that is bought at market rates regardless of which plan you subscribed to.

Put the two halves of the bill side by side and the proportion becomes the real finding. Thirty profiles in the working scenario means roughly US$180 to US$315 a month in traffic. The browser subscription that covers thirty profiles comfortably sits between US$20 and US$49 across the vendors we collected. In other words, on a normal operation the network is roughly four to sixteen times the software, and the industry spends almost all of its comparison energy on the smaller half. Choosing a browser to save ten dollars while ignoring a three-hundred-dollar traffic bill is the most common budgeting mistake in this category, and it survives because only one of those two numbers appears on a pricing page.

One honest limitation of this whole exercise: price per gigabyte says nothing about quality per gigabyte. Two residential pools at the same rate can behave completely differently on the same platform, and the cheaper one is not automatically the worse one or the better one. Which type of proxy fits which operation is a separate decision, mapped in residential, datacenter or mobile.

How AlterAntiX handles your own proxy

We do not sell traffic, which is worth stating plainly in an article like this one: we have no commercial reason to push you toward either side of the comparison above.

Proxies live in their own manager inside the app, separate from profiles. You register one with a name, a type (HTTP, HTTPS or SOCKS5), host, port, optional credentials and a country, and there is a test connection button that reports back the exit IP and the response time in milliseconds before you attach it to anything. The list view shows type, host and port, country, response time and a status column that marks a proxy as active, dead or unknown, with bulk actions to test everything at once and to clear out the dead entries.

Two behaviours matter more than the interface. First, an optional pre-launch check that verifies the proxy answers before the browser opens, so a dead proxy fails as a message instead of as a session that quietly went out through your own connection. Second, the profile fingerprint is synchronised to the proxy country at launch, so the timezone and locale of the profile do not contradict the IP it is about to appear from — the single most common coherence failure in a bring-your-own-proxy setup. Selecting each proxy also shows which profiles are attached to it, which is how you catch the accidental sharing that creates cross-links.

If you want to run the measurement in an environment that shows you the proxy country, the exit IP and the fingerprint coherence in the same place, the desktop build is free — download it here, register one proxy, attach one profile and let it run a normal week. The number you get out is worth more than every price table in this article, because it is the only one that describes your operation.