Proxy Index
OverviewReportsRankings
MethodologyBlogAbout
LensProxy Index
Scoring lensApplied across rankings and provider pages
PresetOur standard score: workload success counts 3×; speed, pool quality, trust, and session count 1× each.For latency-sensitive jobs: speed counts 5×, workload success 3×, and every other dimension 1×.For reputation-sensitive traffic: trust counts 5×, workload success 3×, and every other dimension 1×.For sticky account sessions: session stability counts 5×, workload success and trust 3×, and speed and pool quality 1×.

A score appears only when every dimension with a positive weight has comparable evidence. Saved on this device.

Latest report
Home › Corrections
Public record · reviewed July 30, 2026

Corrections stay visible.

Material errors are corrected in public. Benchmark artifacts remain immutable, prior URLs stay readable, and the current revision explains what it supersedes.

On this page

  • Correction log
  • Materiality and timestamps
  • How correction works
  • Submit a challenge

Related

  • Editorial standards
  • Testing desk
  • Methodology changelog

Public correction log

This log records material publication corrections. Routine new benchmark windows and refreshed measurements are not corrections and do not appear here.

Corrected July 30, 2026

Current scores and ranks were brought under the published scoring contract

Problem. Current ranking and report surfaces treated legacy diagnostic composite values as if they were current Proxy Index scores. The affected publications did not carry the complete current gate, dimension, ranking-context, and sample evidence needed to support those scores or ranks.

Resolution. New immutable publications adopted the current scoring contract and fail closed when evidence is incomplete. The affected last-30-day lines are now diagnostic-only, with no Proxy Index score or rank; the all-time lineage was corrected under the same contract. Historical URLs remain available and point to the corrected current publications.

Affected publications

  • NSTProxy ISP · current last-30-day report · prior July 29 publication
  • Bart ISP static proxies · current last-30-day report · prior July 30 publication
  • NSTProxy ISP · current all-time report · prior July 28 publication

Materiality and timestamps

A correction is material when it changes a conclusion, score, rank, measured fact, attribution, caveat, or a reader's likely interpretation. A material correction receives a dated notice explaining the problem and resolution.

Pure spelling, typography, or formatting can be fixed without a log entry only when the meaning does not change. Methodology changes are dated in the methodology changelog and apply prospectively; historical reports are not silently rescored.

How a correction works

  1. A challenge is checked against the cited source and public evidence.
  2. The testing desk decides whether the issue is factual, interpretive, immaterial, or a change in provider operations that belongs in a new test window.
  3. A benchmark error produces a new immutable publication revision. The historical artifact remains readable and points to the current one.
  4. A material article correction updates the modified date and adds a notice describing what changed.
  5. The public log records the correction without exposing private reporter details.
Provider responsesA response can identify an error or supply context, but it cannot directly change a measurement, score, rank, or placement. Verified errors follow this correction process; operational changes belong in a retest or future benchmark window.

Submit a challenge

A useful submission identifies the affected URL, the exact statement or value, the proposed correction, and a primary source or reproducible evidence. Do not include proxy credentials, endpoint addresses, raw customer data, or other secrets.

Submission route statusThe monitored correction form is not live yet. Proxy Index will not publish an unverified email address or collect sensitive messages until authenticated operator access, retention rules, spam controls, and delivery monitoring are in place.
Read next

Inspect the standard behind the record.

The editorial standard explains evidence, provider responses, automation, conflicts, and the publication checklist.

Editorial standards →Published reports →
Proxy Index

An evidence-first benchmark publication covering residential, ISP, mobile and datacenter proxy providers. Methodology, evidence summaries and benchmark timestamps are published with every issue.

One page per benchmark issue. No tracking pixels.

Reports

  • Latest issue
  • Residential
  • ISP
  • Mobile
  • Datacenter

Methodology

  • Gate & dimensions
  • Measurement nodes
  • Caveat codes
  • Changelog
  • Public data exports

Publication

  • How we're funded
  • Testing desk
  • Editorial standards
  • Corrections
  • Responsible use
  • Blog
© 2026 Proxy Index. Evidence before claims.