Proxy tier benchmark › Targets

Can you scrape Canva template pages with a proxy?

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

No tier tested made this target usable. Unproxied requests were refused outright (0%), the best configuration reached only 40%, and the residential arm — the most expensive — scored 0%. It is also the one target in this study whose success figures should not be trusted in either direction, for the reason below. Measured 2026-08-18.

What was measured
URL https://www.canva.com/templates/
Content assertion Canva — never verified against a live page
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) 0% 149 ms 163 ms n/a blocked 10
datacenter, one pinned exit 10% 766 ms 2,022 ms 8,016 success 1, blocked 9
datacenter, rotating exit 30% 811 ms 1,381 ms 8,372 success 3, blocked 7
anti-detect, one pinned exit 40% 2,524 ms 2,917 ms 925,340 success 4, challenge 6
anti-detect, rotating exit 20% 1,371 ms 2,849 ms 926,536 success 2, challenge 8
residential, rotating exit 0% 2,121 ms 5,321 ms n/a challenge 9, server error 1

What protects this target

Cloudflare, in two different modes depending on the client.

Unproxied and plain-datacenter requests were refused with HTTP 403 carrying a 754,946-754,968 byte Cloudflare branded block page; the reconnaissance probe captured its text: "We'll have you designing again soon. If you can't get past this page, visit the Help Centre...". Requests carrying a browser fingerprint or a residential exit were instead answered HTTP 200 with a 910,665-934,438 byte interactive challenge. That is the study's general pattern in miniature: the more you spend, the further into the defence you get, and none of it converts into a readable page.

Why the numbers here are unsafe. The content assertion for this target is the single word Canva, and it was never verified against a live page (assertion_verified: false in targets-full.json) because the probe was blocked too. Four of the ten recorded successes have a body of 8,016-9,085 bytes; the other six are 923,909-927,319 bytes. Both sizes matched the assertion. An 8 KB response containing the word "Canva" is not the templates listing — the 754 KB block page carries the title Canva as well. So the 10% and 30% figures are probably overstated, and no figure on this page should be read as a pass rate for retrieving template data.

Under sustained production load DIMENSIONS.md measures canva.com failing 42.7% of the time with load spread across many exits and 72.3% concentrated -- it is a hard target at every volume.

robots.txt status

The URL measured is allowed by the host's robots.txt.
Status allowed
File https://www.canva.com/robots.txt
Crawl-delay not declared
Fetched as the benchmark's own User-Agent
Note robots.txt 6324 bytes via benchmark_ua

Limits of this result

  • The content assertion was never verified against a live page, and at least three recorded successes are two orders of magnitude too small to be the requested listing. Treat every success rate on this page as an upper bound with an unknown error.
  • The templates index was measured, not an individual template page.
  • 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 '^[^,]*,canva_template,' 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