About SpeedTestHQ
SpeedTestHQ is a free, browser-based internet speed test that runs directly in your browser. We measure download, upload, ping, and jitter against a global network of test endpoints, then publish guides and benchmarks built on the patterns we see in real-world connections.
How the Speed Test Works
Our test runs four independent measurements in your browser, with no software to install:
- Download: Multiple parallel HTTPS streams from the closest edge server, ramped to saturate your line. We measure throughput over the steady-state portion of the transfer (after the TCP slow-start ramp) and report the median.
- Upload: Parallel POST streams using the same approach, measuring sustained throughput once your line reaches steady state.
- Ping (Latency): Round-trip time of small HTTP requests to the test server. We send 20 probes and report the median to filter out one-off network jitter.
- Jitter: The mean absolute deviation of ping samples. High jitter is what makes video calls and VoIP sound choppy even when the headline ping looks fine.
Methodology Details
- Server selection: Tests automatically connect to the closest available edge — this is what your everyday browsing actually traverses, so results match what you experience day to day.
- Steady-state measurement: Short bursts overstate speed by capturing TCP slow-start. We discard the ramp-up window before calculating throughput.
- Parallel streams: Single-stream tests systematically underreport on connections above ~500 Mbps because of TCP per-stream limits. We use 4–8 parallel connections to remove that bottleneck.
- Browser overhead: Some throughput is lost to TLS, browser buffer copies, and the JavaScript event loop. On gigabit connections this can account for a 5–10% gap between our number and what a wired iperf3 test would show.
Where Our Benchmark Data Comes From
The state, country, and ISP averages published in our reports are derived from aggregated, anonymized speed-test results combined with publicly available regulatory data (FCC Form 477 in the US, Ofcom data in the UK, ANATEL in Brazil, and similar sources for other markets). We cross-check our aggregates against published ISP plan tiers and third-party datasets such as M-Lab and Ookla's open data releases to flag outliers.
Numbers are refreshed periodically. Each report page lists its data window and the methodology used to derive averages, percentile ranges, and ISP rankings.
What We Don't Do
- No personal speed-test profiles. We don't build an identity-based history of your individual speed-test results. Advertising and engagement technologies may process page-view and device data as described in our privacy policy.
- No paid placements in rankings. Our "best" lists and ISP rankings are not pay-to-play. Affiliate disclosures, where any apply, are noted on the page in question.
- No fake "your speed is slow" upsells. The test reports your numbers and what they mean for common activities — that's it.
Data Practices
We minimize the speed-test data we keep. Test results are stored in aggregate for benchmarks; we do not retain a personal history of individual results tied to your IP. Some pages may include advertising, analytics, or engagement technologies from third-party providers. Full details are in our privacy policy.
Contact & Corrections
Spotted an error in a benchmark, an outdated ISP plan, or a broken link? Let us know via the contact form. We update reports and guides as conditions change — corrections are usually applied within a week of being verified.