How we test

Every page on this site is built on three rules: a figure is either read from the provider’s own page on the date shown, or observed on a machine we ran ourselves, or it is marked not published. Nothing is estimated, nothing is recycled from another site’s tests, and every change is logged. This page is the procedure, so you can check a page against it.

What we test, and how

• Real machines, real sessions. Antidetect browsers are installed and run on real hardware: we create profiles, launch them, and check what the product actually exposes — profile limits, fingerprint fields, proxy binding, team seats, API access. Proxy providers are tested by running real requests through real sessions: rotation behaviour, sticky-session duration, targeting options and failure handling are observed, not copied from marketing pages.
• We do not test at scale. We run a small, reproducible suite, not a 10,000-request benchmark. Anything we did not measure ourselves is not presented as a measurement; where a vendor publishes a claim we could not verify (pool size, success rate, response time), the page says it is a vendor claim and stays dated.
• Screenshots and logs where they help. When a limit or a behaviour is central to a verdict, the page links the evidence. What we cannot show, we do not assert.

How we price

• Every price is read from the vendor’s own pricing, documentation or legal page — never from an aggregator, a review roundup or a blog post — and published with the date it was read. The date is printed on the page and in every row of the dataset.
• A price is what the vendor publishes, with the term it requires. A promotional rate is marked as promotional; a rate that requires annual billing is marked as annual; a renewal price is checked and published if it exists, and marked not published if it does not.
• Currencies are not converted. Each figure stays in the currency the vendor publishes it in. Converting would introduce an exchange rate that is our choice, not the vendor’s fact.

Status markers

• Pages and rows carry status badges: verified (re-read on the date shown and still current), expired (the offer or price is gone), live (offer currently offered), updated (a figure changed and was re-dated).
• Markers are declared, never inferred. We do not stamp a row verified because a price looks plausible; we re-read the page, and the badge carries the date of that re-reading.

The changelog

• Every change to a figure is recorded in the Data hub changelog: what changed, when, and which vendor page the new figure was read from.
• An old value is never silently kept as current. When a provider changes a price, the row is re-read, re-dated, and the old value is superseded — the changelog is where the history lives.

What we do not do

• No vendor review copies, no sponsored rows, no paid placement. A provider cannot buy a row, a score or a position.
• No “best of” lists built from other people’s tests. If we did not read it from a first-party source or run it here, it is not published.
• No invented figures. If a provider does not publish something we need, the cell says not published — that is a finding, not a gap.

Who writes this

Runein is written and maintained by the site operator — an independent developer who pays for and deploys the tools these pages cover. There is no editorial team, no newsroom and no sponsored section. Corrections go to [email protected] and take priority: send the page URL, the row, and the first-party source that shows the current value, and the figure is re-read, re-dated and logged.