Proxy tier benchmark Targets

Do you need a proxy to scrape AliExpress product pages?

High-defence commercial sites · minimum viable tier: no proxy needed · measured 2026-08-18

Not at this volume, and if you do use one, turn Anti-Detect Mode on. An AliExpress item page returned the expected content on 10 of 10 unproxied requests (100%). Adding a plain datacenter proxy made it worse — 60% on a pinned exit — and enabling Anti-Detect Mode over the same datacenter pool restored it to 100%. Measured 2026-08-18.

What was measured

URL https://www.aliexpress.com/item/1005005167379524.html
Content assertion Aliexpress
Arms 6
Rounds per arm 10
Requests 60
Measured 2026-08-18, 15:38–18:09 UTC
Benchmark host DigitalOcean droplet, New York (nyc3)

A request counts as a success only if the expected content is present. An HTTP 200 carrying a challenge page, an empty body or the wrong page is a failure, because it is not a page you can use.

This measures whether a tier works, not long-run stability under sustained load. Ten requests per arm, at a two-second floor, never approach a per-IP rate limit — which is the thing a rotating pool exists to defeat.

Results by arm

Arm Success p50 p95 Bytes per success Outcomes
direct (no proxy) 100% 435 ms 1,435 ms 77,606 success 10
datacenter, one pinned exit 60% 672 ms 2,282 ms 77,606 success 6, challenge 4
datacenter, rotating exit 80% 1,416 ms 2,644 ms 77,606 success 8, challenge 2
anti-detect, one pinned exit 100% 1,423 ms 3,831 ms 77,606 success 10
anti-detect, rotating exit 90% 936 ms 4,493 ms 77,606 success 9, challenge 1
residential, rotating exit 100% 3,504 ms 4,450 ms 77,606 success 10

What protects this target

A captcha interstitial served at HTTP 200. Seven of 60 requests came back 200 with a 2,337-2,379 byte body matching the harness's generic captcha signature, against 77,606 bytes for the real item page. All seven landed on proxied arms that were not anti-detect on a pinned exit: four on the pinned datacenter exit, two on the rotating datacenter exit, one on anti-detect rotating. Neither the unproxied arm nor anti-detect on a pinned exit saw a single one.

Two things follow. First, the trigger is not only the IP address — the anti-detect arms drew from the same datacenter pool, so the signal that fired was the TLS and HTTP/2 fingerprint of a plain Python client. Second, a harness scoring HTTP status codes would have recorded 60 of 60 successes here and published 100% for every arm.

Under sustained load the picture changes again: DIMENSIONS.md measures aliexpress.com failing 8.9% of the time when a workload is spread across many exits and 32.7% when it is concentrated on few.

robots.txt status

The URL measured is allowed by the host's robots.txt.

Status allowed
File https://www.aliexpress.com/robots.txt
Crawl-delay not declared
Fetched as the benchmark's own User-Agent
Note robots.txt 1971 bytes via benchmark_ua

Practical setup

If you proxy this target, enable Anti-Detect Mode — on the measured pages it was the difference between 60% and 100% over the same exit pool. Spread requests across exits with a fresh session id per request rather than pinning one.

Limits of this result

  • One item page was measured. AliExpress serves different defences to search, category and item URLs, and this result describes an item URL.
  • The 8.9% / 32.7% production figures are derived from response-size distributions rather than status codes.
  • Ten rounds per arm in a single three-hour window on 2026-08-18. Anti-bot behaviour drifts; a result older than a few months is a historical record, not a current one.
  • One exit geography. Every request left from a US datacenter host or a US exit, so nothing here describes how this target answers traffic from elsewhere.
  • The unproxied arm ran from a clean, previously unused cloud IP with no accumulated reputation. It is a best case a production scraper will not sustain.

Run date: 2026-08-18 · Full benchmark report · Harness and raw data · All 32 targets

The 60 rows behind this page are in results/six-arm/raw.csv; filter them with grep '^[^,]*,aliexpress_item,' raw.csv.


Test your own target

The result above describes one URL on one day. Paste your own URL into the live tester and see the real response before creating an account.

Test your target See plans