NETWORK / DIAGNOSTICS
Test your connection.
Download. Upload. Latency. A clear reading of your internet.
CONNECTION / LIVE
Ready when you are.
Loading the measurement engine...
Target: Cloudflare's network. HTTP measurements from this browser, not a promise of your plan's maximum speed. Packet loss is not measured.
Start only when you're ready to use test data. Your IP address is visible to Cloudflare when testing. Cloudflare privacy. Bremo saves results on this device, not in a public database. Basic page analytics are separate.
RESULT / SUMMARY
Your connection, explained.
HISTORY / LOCAL
Your recent tests.
Up to 10 completed tests, saved in this browser. Try Ethernet, then Wi-Fi, or compare different rooms.
| When | Profile | Download | Upload | Ping | Jitter |
|---|
Methodology & test notes Read how measurements work
BEFORE YOU PRESS GO
Give your connection
a fair test.
Pause large downloads and cloud backups. For a baseline, connect with Ethernet. Then test over Wi-Fi where you actually use it.
- Keep the tab open.Switching tabs or locking the screen can distort a browser test. We'll stop an active test if the page goes into the background.
- Choose your data budget.Use Light on a mobile hotspot or metered plan. Extended uses larger transfers to examine faster connections.
- Repeat when it matters.Try morning and evening. One result is a snapshot, not an assessment of your provider's whole network.
HOW IT WORKS
Real transfers. Clear limits.
What we measure
We use Cloudflare's open-source measurement engine, served by Bremo, to transfer test data directly between your browser and Cloudflare. Download and upload use increasing payload sizes. Ping is HTTP round-trip latency, and jitter reflects variation between consecutive latency samples.
Bandwidth is calculated from browser Resource Timing data. The result uses the engine's 90th-percentile throughput and median latency. This is an adaptive browser transfer test, not Ookla's engine or server network.
What changes the result
Wi-Fi interference, your device, VPNs, browser extensions, other traffic and the route to the measurement server all affect results. Lightweight tests can understate very fast connections. Loaded ping is shown only when a transfer lasted long enough to collect usable samples.
We do not infer your ISP, precise location or packet loss. Our use-case tips are simple heuristics, not a certification. No speed or performance score is invented when a measurement is missing.
Measurement engine and methodologyOpen-source licenceBremo privacy
What speed do I need?
Needs vary with devices and applications. Our quick tips use 25 Mbps download as a starting point for one high-resolution stream, and 10 Mbps download plus 5 Mbps upload for a video call. Gaming tips consider measured latency and jitter as well as throughput. Packet loss is not tested, so no gaming result is a guarantee.
Why is this different from another speed test?
Tests can use different server locations, protocols, parallel connections, payload sizes and calculations. Compare repeated results from the same tool and setup. A single test cannot establish your service plan, diagnose every fault or prove a provider is delivering its advertised maximum.
Where are my results stored?
The last ten complete results are stored in this browser when local storage is available. Bremo does not send this history to a results server. We disable the measurement engine's optional results-logging endpoints. Cloudflare still processes your test traffic under its own privacy policy, and Bremo's basic site analytics record page activity.