iSpeedTestA speed test that helps you take the next step
A speed test should do more than hand you a big number. It should help you decide whether the connection in front of you is suitable for a video call, a film, a game, or a large upload. Start the test above, leave this tab in the foreground, and wait for the individual measurements to finish. You will get separate download and upload rates in megabits per second, together with HTTP latency and jitter in milliseconds.
This speed test transfers generated data between your browser and Cloudflare measurement endpoints. It measures that particular path at that particular moment. It does not certify your broadband plan or establish the performance of every website. The result becomes more useful when you repeat the same experiment under controlled conditions and change one thing at a time. Start with the device and room where you actually experience the problem.
How to run a more useful speed test
Before a baseline speed test, pause large downloads, cloud backups, game updates, and streaming on other devices. Keep the browser visible because background tabs can be throttled. If possible, plug a computer into your router with Ethernet and record that result first. Then repeat over Wi-Fi in your usual room. The difference helps separate an internet access issue from a wireless coverage issue.
Run another speed test later in the evening if that is when things feel slow. Record the time, connection type, and whether a VPN was active. A single unusually low reading is a reason to investigate, not proof that your provider has failed. Repeated measurements under matching conditions are a much better basis for a support conversation. On a limited mobile plan, select Data saver before starting. Testing uses real data, and running many tests can consume a meaningful part of your allowance.
Understand the four numbers in your speed test
Download measures how quickly this browser receives the test payload. It matters for streaming, downloading files, and installing games. Upload measures how quickly it sends a generated payload; video calls, cloud backups, and sending large attachments all depend on it. Our speed test calculates throughput from transferred bytes and elapsed request time, including browser and HTTP overhead. The two directions can differ substantially, especially on asymmetric plans.
Latency describes the time for an HTTP request and response. Lower is generally better for responsiveness, but a browser measurement includes more than network propagation alone. Jitter describes variation between consecutive latency samples. A connection can have a fast download result and still feel inconsistent if response times fluctuate. The speed test also samples latency while transfers are active when samples are available. Missing measurements stay unavailable; they are never filled with zero or invented estimates.
What does a good speed test result look like?
A useful connection has enough capacity for the things happening at the same time, plus predictable response times. Netflix recommends at least 5 Mbps for a single Full HD stream and 15 Mbps for a single 4K stream. Those figures are application recommendations, not a complete household budget. A second screen, software updates, and remote work all share available capacity. Leave headroom instead of matching a plan to the absolute minimum.
For gaming, a speed test is only part of the story. In-game latency depends on the game server and route, which may differ from this measurement endpoint. For remote work, upload capacity and stable latency often deserve as much attention as download capacity. Use the guides below to interpret the task that matters to you. Avoid buying a faster plan solely because somebody else's speed test shows a larger number; first identify the actual bottleneck.
Your speed test on Wi-Fi, fiber, and mobile
A Wi-Fi speed test includes both the local wireless link and the internet connection beyond the router. Walls, interference, distance, mesh backhaul, and device radios can all change the result. It cannot directly report your Wi-Fi signal strength or the negotiated radio link rate. Moving near the router is a useful comparison, but it does not reproduce the experience at your desk or television.
Ethernet removes many wireless variables, although an old cable, adapter, or port can still become a bottleneck. Fiber does not eliminate device limitations or congestion elsewhere on a route. Mobile performance changes with signal conditions, cell capacity, building materials, and whether the device is using LTE or 5G. Make a speed test part of a repeatable comparison: same device, same location, same endpoint, and a written note of the connection being used.
Why your speed test results can change
Internet measurements are observations, not fixed properties of a subscription. Other people on your network, router processing, browser activity, server load, and congestion between networks can affect a run. Different testing tools can use different destinations, file sizes, connection counts, and calculations. Two tools can therefore return different results without either result being fabricated.
This speed test uses bounded HTTP transfers and is intended as a practical browser check. It may under-read very fast links, particularly when a short transfer spends a large fraction of its time starting up. A native testing application or a longer controlled transfer can provide another perspective. If a result seems implausible, check the selected profile and compare several runs. Do not average together tests taken on unrelated networks and treat the combined number as a meaningful diagnosis.
Turn a slow speed test into a troubleshooting plan
First compare Ethernet with Wi-Fi. If Ethernet is healthy but wireless is slow, work on placement, interference, channel selection, or mesh connectivity. If both are slow, check the subscribed plan, router ports, background traffic, and provider service information. Restarting a router can resolve temporary problems, but repeated restarts erase useful evidence and will not fix every fault.
Use a speed test before and after each change. Keep the conditions as similar as possible. If latency rises sharply during uploads, pause backups and investigate queue management on the router. If only one application is slow, test that application's own diagnostics rather than assuming the whole connection is broken. When contacting support, provide several dated readings, your connection method, and the observed symptom. A short, consistent test record is more actionable than a screenshot with no context.
Designed for a deliberate, transparent speed test
No test starts until you press the button. The standard profile is capped at 96 MB of test payload across both directions, while Data saver is capped at 12 MB; protocol overhead is additional. Slower connections may use substantially less. You can stop a running speed test, and incomplete results are clearly identified rather than presented as a completed run.
Recent results are saved on this device only if you enable the history option. You can clear them whenever you like. Cloudflare receives the network requests and your IP address as the measurement provider. Read the methodology and privacy pages for the exact scope. We do not claim to be your provider, to operate a global server network, or to guarantee a particular result. A trustworthy speed test makes its limitations visible alongside its useful measurements.
Sources & further reading
Our explanations distinguish the browser measurement from the underlying connection. These primary references provide context; they do not endorse this site.
Cloudflare: browser measurement methods Netflix: connection recommendations for streaming MDN: browser resource timingLet’s clear things up.
Is this speed test free?
Yes. You can run the browser test without an account or payment. Your internet or mobile provider may still count the transferred data against your allowance.
Does a speed test measure my Wi-Fi signal?
It measures internet performance through your current connection. It cannot read radio signal strength or isolate local Wi-Fi capacity from the internet path.
Why is my speed test lower than my plan?
Device limits, Wi-Fi conditions, background traffic, routing, and the bounded test methodology can all contribute. Compare a wired baseline before drawing conclusions.
Can I compare a speed test with Ookla or FAST.com?
Yes, as separate observations. Different endpoints and measurement methods mean the numbers are not interchangeable. Record the tool and conditions with each result.