ISP static proxiesmethodology Canonical proxy scorecard v1Revision 68

Bart ISP · ISP static proxies

Bart ISP static proxies last 30 days benchmark passed. Proxy Index comparison score 99.6/100.

Inspect report evidence ↓Read methodology Canonical proxy scorecard v1 →

Provider tested

975

Network checks

Worker regions

Evidence families

Last 30 days · Sep 6, 2026 · Revision 68 · Published by Proxy Index testing desk · Editorial standards

Sample designLast 30 days

Grades are measured from us-east. Also observed, reported for context and never graded: eu-west. Declared concurrency tiers: 6-20.

Metric coverage32/36 measured

Missing metrics are listed explicitly instead of implied.

Measurements from this immutable publication

These dimensions and inputs belong to Bart ISP's isp static proxies report. Current cohort ranks are intentionally kept separate from this report record.

Benchmarkablepassed 6/6 baseline checks

Connection
Exit identity
HTTP delivery
ASN lookup
Geo lookup
Throughput

Workload success100

Mean of equal-weighted workload-fixture success rates, scaled to 0-100. Each fixture already includes its configured raw HTTP and browser checks, so browser results are not counted twice. Baseline plumbing checks do not raise this grade.

Availability heartbeat success rate 100% · E-commerce product fixture success rate 100% · JSON workload fixture success rate 100% · Large HTML fixture success rate 100% · Listing page fixture success rate 100% · SERP fixture success rate 100%

Speed100

Latency policy proxy_index_speed_latency_log_v2: raw HTTP requests through the proxy, successful attempts, a fresh connection every time with setup included, against three named targets (our fixtures, a CDN edge, a fixed European origin). Each target's median and p95 are scored on one absolute log curve (100 at <=100ms, 0 at >=10000ms) and blended 50/50; component scores combine with equal weight over the components measured this window (fixtures). Averaged with throughput graded as min(100, Mbps x 10) when present.

Request latency p50 31 ms · Request latency p95 41 ms · Throughput 93 Mbps

Pool quality97.2

Static ISP pool grade: mean of endpoint-to-exit coverage, repeat stability, exit distribution balance, and exit-IP entropy. Stable reuse is expected for a static endpoint and is not penalized as failed rotation.

Unique exit IP count 25 · Repeat rate 88.9% · Top exit IP share 4% · Exit IP entropy 1

Trust and sourcing100

Geometric mean of enrichment-derived trust signals scaled to 0-100: usage-type accuracy (claimed vs observed), the inverse of high-risk IP share, and country geo accuracy. A zero load-bearing signal cannot be averaged away. Enrichment inputs require at least five unique exits, at least 80% coverage, and one labelled vendor source.

Usage type accuracy 100% · High-risk IP share 0%

Session100

Sticky-session hold rate scaled to 0-100. Requires session-persistence attempts.

Session hold rate 100%

Value—

Not graded: Price is context by design. The registry does not grade it.

The evidence

The measurements behind the dimension grades above: what we ran and what it returned.

Attempt accounting

Checks passed975 of 975

Completed attempts975

Raw failed attempts0

0 Proxy Index infrastructure failures excluded and reported separately.

Reliability

Scrapers need requests to complete before speed, targeting, or browser behavior matters.

Connection100HTTP reachability100Success rate100%

measured

Speed

Latency and transfer speed determine queue depth, worker cost, and retry pressure.

Throughput score100Request latency p5031 msRequest latency p9541 msEdge latency p50n/a

partial

Pool quality

Rotating proxy buyers need a pool that is broad enough and not concentrated on a few exits.

Exit identity score100Unique exit IP count25Repeat rate88.9%Pool churn rate0%

measured

Target fit

HTTP 200 is not enough; buyers need to know whether representative pages return usable content.

Average target success rate100%Average semantic success rate100%Average block rate0%Average challenge rate0%

measured

Browser readiness

Browser-based scraping depends on proxies surviving the same engine stack buyers use.

Browser engine success rate100%Browser p50 latency118 msBrowser p95 latency142 msCAPTCHA rate0%

measured

Metadata accuracy

Regional and network targeting breaks when provider metadata is stale or incorrect.

Country accuracy score100Region accuracyn/aCity accuracyn/aASN accuracy score100

partial

Metric coverage

32 of 36 tracked buyer metrics are measured in this report.

MeasuredSuccess rate
MeasuredBrowser engine success rate
MeasuredBrowser p50 latency
MeasuredBrowser p95 latency
MeasuredP50 latency
MeasuredP95 latency
MeasuredVantage-region p50 latency
MeasuredVantage-region p95 latency

Target test outcomes

What happened when this proxy handled representative scraping targets. This is direct evidence behind the workload success dimension, not a buying recommendation.

Met test thresholdAvailability heartbeatObserved 100.0% success · 412 ms p95 latency · 0.0% blocked in this published sample.
Met test thresholdProduct pageObserved 100.0% success · 132 ms p95 latency · 0.0% blocked in this published sample.
Met test thresholdJSON APIObserved 100.0% success · 143 ms p95 latency · 0.0% blocked in this published sample.
Met test thresholdHeavy HTML pageObserved 100.0% success · 150 ms p95 latency · 0.0% blocked in this published sample.
Met test thresholdCategory listingObserved 100.0% success · 128 ms p95 latency · 0.0% blocked in this published sample.

Recent time series

6 public trend series are available for this report.

n/aASN accuracy3 data points
0%Benchmark infrastructure failure rate3 data points
0%Block rate3 data points
117 ms–120 msBrowser p50 latency3 data points
139 ms–142 msBrowser p95 latency3 data points
100%Browser engine success rate3 data points

How this issue was scored, and the fine print

Every number above comes from the published report for Last 30 days, with 32 of 36 buyer metrics measured. The reference matter and citation for this issue follow. Machine-readable exports are published for the whole feed, not per issue. See public data exports.

Commercial contextCommercial relationships cannot alter measurements or rank. Any material provider relationship affecting this publication must be disclosed with the report.

Notes on this window

Incomplete daily coverage · The rolling window has fewer daily data buckets than calendar days.
Mixed vantage regions · This report observed multiple worker vantage regions. Grades are measured from us-east only; the other regions publish through the vantage-region success and latency series and never enter a grade.

Ranking eligibilityRanking eligible means the report passed the public gate, dimension, sample, confidence, and coverage requirements and may enter the current rank; it is not a product recommendation. It observed 25 endpoints against a 5-endpoint minimum and 3 eligible runs against a 2-run minimum, at 100% confidence. Meets the public gate, comparison, and sample thresholds.

What the published report carries

The published report carries the eligibility gate, buyer dimensions, metric definitions, sample design, per-page results, and the redaction policy. It withholds credentials, endpoint hosts, raw observations, and raw attempt errors.

Redacted fields: endpoint.id, endpoint.host, endpoint.port, endpoint.proxySourceId, endpoint.tags, proxy credentials, raw observations, raw attempt errors

Follow the evidence past this issue

This issue is one window on the field. Open a provider's standing record for the full history, or read the methodology behind every number.

Open Bart's record →Read the methodology →