Real Time & Accurate Internet Speed Test Online

What this page measures

The test runs three phases and reports four numbers. Ping is the round-trip time to the test server in milliseconds. Jitter is how much that time varies between requests. Download and upload are measured in megabits per second while the browser transfers data for a fixed window.

The result reflects the path between your browser and the selected server at that moment. It is not a permanent property of your connection, and it is not a guarantee from your internet provider.

Ping and jitter

The page sends eight small requests to the server and times each one. It sorts the samples, drops the fastest and slowest, and averages the rest. Dropping the extremes reduces the effect of a single scheduling hiccup. Jitter is the average absolute difference between each remaining sample and that average.

Lower is better for both. A low average with high jitter means the connection is inconsistent, which can hurt calls and games even when the headline speed looks fine.

Download speed

The download phase runs for eight seconds and uses four parallel streams. With the local server, each stream requests a 10 MB block; with a CDN, each stream requests a file from the selected public CDN. Bytes are counted as they arrive and the speed is recalculated during the run.

Parallel streams are used because a single HTTP connection often cannot fill a fast link. The final number is total bits divided by elapsed time, converted to Mbps.

Upload speed

The upload phase runs for six seconds and uses three parallel workers. Each worker posts 256 KB of generated data in a loop and counts completed bytes. Upload testing only works with the local server, because the public CDNs used here accept downloads but do not provide an upload endpoint.

If you select a CDN, the page still tests ping and download, but it skips upload and the upload card stays empty. Choose "This Server (full test)" when you need all four metrics.

Choosing a server

"This Server" tests against the machine that serves this site and supports ping, download and upload. "Best Server" chooses automatically from the available list. The jsDelivr, unpkg and cdnjs options test download through a public CDN and are marked download-only.

Distance matters. A nearby server measures your access link more cleanly; a distant server includes more internet routing. CDN results can be flattering because the CDN may have a cache close to you. Use the same server option when comparing two runs.

Client information

The page makes a request to a public IP lookup service to show your approximate network and provider. That request goes to a third party, not to this tool's server. If your browser blocks it, the test still runs; only the client information line is missing.

If you want to inspect your browser and device instead of your connection, the What Is My Device page covers that separately.

Reading the results

Mbps means megabits per second. To convert to megabytes per second, divide by eight. A 100 Mbps result is about 12.5 MB/s, which is the number a download manager usually shows. Ping and jitter are already in milliseconds.

There is no universal pass mark. Video calls care about latency and jitter more than raw download speed. Large downloads care about throughput. Gaming cares about latency and stability. Compare your result with what your plan advertises and with an earlier test on the same server.

Why results vary

Wi-Fi signal, distance from the router, interference, the device's network adapter, browser limits, VPNs, other downloads, background updates and the time of day can all change the number. A laptop on Wi-Fi in another room will usually test lower than a wired desktop.

Keep the tab in the foreground while the test runs. Browsers may throttle timers and network work in background tabs, which can distort the live chart and the final average.

What it does not measure

It does not measure packet loss, bufferbloat under load, Wi-Fi coverage, per-app priority, or long-run stability. It does not run a multi-server average or a traceroute. A single result is a snapshot, not a full diagnosis of an intermittent fault.

If you are troubleshooting a problem, run the test several times at different times of day and on different devices before contacting your provider. Note the server option, the time and whether you were on Wi-Fi or Ethernet.

Improving a poor result

Test again with a wired connection if you can. Move closer to the router or switch to the 5 GHz band, close background downloads and pause cloud backups. Turn off the VPN and any proxy, because both add a hop and can cap throughput. Try a different server option and a different time of day; a single result at peak hour is not enough to conclude that the line is faulty.

If every device on the network tests slowly at the same time, the problem is likely the line or the router. If only one device is slow, start with that device's Wi-Fi adapter, drivers and background apps. Record the server, the time, the connection type and the numbers before contacting your provider, so the support agent has something specific to work with.

Step by step

1. Close large downloads and pause streaming on other devices if you want a cleaner result.
2. Choose This Server for a full test, or a CDN for download-only.
3. Press Start Test and keep the tab visible.
4. Watch the phase indicators and the live gauge.
5. Read ping and jitter first, then download and upload.
6. Press Test Again to compare, using the same server option.

Privacy and limits

The test sends requests to this site's server or to the CDN you select, and an IP lookup request to a third-party service. The tool does not store your results or require an account. There is no export or share button; note the numbers yourself if you need a record.

The test is browser-based and depends on the selected server, the number of parallel streams and the device. It is not an official measurement from your internet provider and should not be used as the only evidence in a service dispute.

Frequently Asked Questions