An open-source harness for measuring real-world Time to First Byte (TTFB) of web hosting from real consumer connections across Australia: Sydney, Melbourne, Brisbane and Perth.
It powers the independent benchmarks published at Pick a Host (how we test), and is released so anyone can reproduce or audit the numbers.
Most "fastest hosting" claims are measured from a single city, or worse, from a CDN edge that hides the origin server entirely. On one Cloudflare-fronted host we measured 5 ms from Melbourne (edge cache) vs 939 ms from Sydney (origin) for the same URL. A single-location test would have called that host blazing fast. It isn't.
This tool measures from multiple real networks per city and tests the bare origin (caching off) so the result reflects the server, not a cache layer.
- Uses Globalping's free distributed probe network (250 tests/hour unauthenticated) to run HTTP measurements from real consumer/ISP probes in each Australian capital.
TTFB = result.timings.firstByte(ms), aggregated across probes per city.- Tests a byte-identical bare-origin page on each host so every target is compared like-for-like.
# Benchmark one URL ad hoc
python3 run-benchmark.py --target https://example.com/ --dry-run # preview probes
python3 run-benchmark.py --target https://example.com/
# Build the published outputs (table + chart dataset) from results
python3 build-outputs.pyOptional: set GLOBALPING_TOKEN for the higher 500 tests/hour limit. No key is required to run.
- Bare origin only. Hosts behind a CDN are tested at the origin with caching disabled, or the result measures the CDN, not the host.
- Real probes, not synthetic. Globalping probes run on real consumer/ISP connections, so TTFB reflects what an Australian visitor actually experiences.
- All four mainland capitals. Sydney, Melbourne, Brisbane and Perth, because east-coast-only testing misrepresents WA and QLD visitors.
MIT. Maintained by Pick a Host - independent Australian web hosting reviews, tested and compared.