MyIP.dogChecking…
Internet speed test
Measure download, upload and HTTP latency from this browser to Cloudflare. Choose a data budget before starting.
Checking…
Measure download, upload and HTTP latency from this browser to Cloudflare. Choose a data budget before starting.
Checking…
Checking…
Larger samples for faster connections. Up to 167 MB of test payload, plus network overhead. Stops after 60 seconds.
Starting sends test traffic and your IP address directly to Cloudflare. Results stay in this page unless you copy them. Switching tabs stops the test. Provider information
No test traffic has been sent.
An internet speed test helps you investigate slow downloads, delayed uploads and inconsistent browsing. Start with the question you want to answer: whether this device can transfer data reliably, whether Wi-Fi placement makes a difference, or whether your VPN changes the path. The numbers describe this browser’s connection to a specific measurement service at a particular time.
This internet speed test sends traffic directly between your browser and Cloudflare. It measures download throughput, upload throughput and HTTP response timing. MyIP.dog calculates the displayed results locally from completed requests. A test does not begin when this page opens; you choose a mode and press Start speed test.
Before an internet speed test, pause large downloads, cloud backups and other tests on your own devices if you want a quieter baseline. Keep the same device, browser and connection when comparing runs. If normal household activity is what you want to investigate, leave it running and record that context alongside the result.
The standard internet speed test requests progressively larger payloads, up to 166.6 MB across both directions. The selector rounds that allowance to 167 MB. Protocol overhead, retransmissions and small responses add traffic, so this figure is not a mobile billing cap. The test stops after sixty seconds or when you stop it.
The data saver internet speed test requests up to 6.6 MB of transfer payload, shown as a 7 MB allowance. Smaller transfers are useful on a metered connection, but they can underestimate a fast connection. Choose the standard mode when you need larger samples and your data plan comfortably allows them.
An internet speed test can compete with a video call or another person’s download. Run it at a suitable moment. Stop test interrupts active browser requests and prevents further samples. Switching to another tab also stops collection; clearing the result removes the displayed measurements without starting a replacement test.
Download in this internet speed test describes how quickly completed response payloads reached the browser, expressed in megabits per second. A larger number means more bits per second for those measured requests. It does not guarantee that a different website, file server or streaming service will deliver data at the same rate.
Upload in this internet speed test describes outgoing test payloads divided by their measured HTTP request duration. The duration includes the wait for the service’s response, rather than pretending that the browser can observe every intermediate packet. Cloud storage, live broadcasts and large attachments can depend on upload performance as well as their own service limits.
The internet speed test uses decimal units: one Mbps is one million bits per second, and one MB is one million bytes. Eight bits make one byte. A result of 80 Mbps corresponds to a theoretical 10 MB per second before accounting for application behavior and other overhead; it is not a promise about a particular file download.
Idle HTTP latency in this internet speed test is the median time from a measurement request starting to its first response byte. An initial warmup request is excluded, followed by ten observations when the run completes. Server processing remains included. This timing is related to responsiveness, but it is not an ICMP echo measurement.
HTTP jitter in this internet speed test is the mean absolute difference between consecutive idle latency observations. For samples of 10, 20 and 5 milliseconds, the adjacent differences are 10 and 15, giving 12.5 milliseconds. At least two observations are necessary. Missing samples cannot establish zero variation.
A completed internet speed test can also show HTTP latency collected concurrently with downloading or uploading. These additional requests begin only while transfer stages are active. Small transfers can finish before a concurrent observation completes, leaving that value unavailable. Such a gap is an absence of evidence, not a finding that the connection has no delay.
Use an internet speed test to look for repeatable changes rather than one universal pass mark. If concurrent latency is regularly higher than idle latency, shared capacity or queueing deserves investigation. This tool does not prove complete link saturation, identify the router responsible, or assign a bufferbloat grade from a few samples.
The internet speed test shows the number of usable observations next to each summary. Throughput samples shorter than ten milliseconds are excluded because a very short timing interval can make the rate unstable. If no eligible sample remains, the result says Not measured. It does not substitute the speed advertised in your browser or internet plan.
A partial internet speed test keeps the measurements that actually finished. An interrupted download can leave upload unmeasured, while a per-request timeout can end one transfer stage and allow the other direction to proceed. Read the outcome and any explanation before comparing these numbers with a completed run.
An internet speed test that reaches a rate limit stops without automatically retrying. A failed request can also reflect a blocked measurement domain, browser restriction, network interruption or service problem. It is not sufficient evidence that your entire internet connection is offline. Wait before rerunning, and check whether other ordinary websites still load.
The internet speed test first warms the connection, then measures idle HTTP timing, download and upload. Standard transfer sizes progress through 100 KB, 1 MB, 10 MB and 25 MB. The first three sizes permit three requests each; the final size permits two. The smaller mode stops its size sequence at 1 MB.
For each direction, the internet speed test stops increasing payload size once every request in a completed size group takes at least one second. This avoids forcing a slow connection to download the largest file. An individual request has an eight-second deadline, and the complete run has a sixty-second deadline, so unfinished work remains bounded in time.
The internet speed test reads network timing exposed through the browser’s Resource Timing interface. Transfer duration runs from request start to response completion. Download bytes are checked against the response body and its timing metadata. Cached, compressed, mismatched or otherwise unusable download observations are excluded instead of generating an impressive but unsupported rate.
The internet speed test summarizes eligible transfer rates using a linearly interpolated ninetieth percentile. This is a defined summary of the observed samples, not a maximum theoretical bandwidth calculation. You can expand Measurement details and samples to inspect the individual payload sizes, HTTP durations, rates and exposed protocol values behind it.
This internet speed test uses sequential payload transfers. Some other tools open multiple transfers to fill a connection more aggressively. Short samples, HTTP request overhead, connection reuse and different destinations can therefore produce different results. This implementation does not subtract an estimated server delay or adjust the answer until it matches a purchased plan.
Your internet speed test result includes the selected mode, measurement endpoint, timestamps and elapsed time. Requested payload includes transfers started but not necessarily completed. Completed payload counts received downloads and uploads acknowledged by a successful response. Neither total is a packet capture or a reliable replacement for your carrier’s data meter.
Run an internet speed test close to the router and then in the room where problems occur, using the same device and mode. Walls, interference and shared airtime can change wireless performance. Repeat each situation a small number of times at comparable moments. A consistent difference supports further Wi-Fi investigation, while one isolated result may reflect unrelated traffic.
If the internet speed test improves on a wired connection, that comparison helps separate the wireless segment from the rest of the route. It still does not identify a particular faulty component. Check negotiated link speed, cable condition and router settings alongside the measurements rather than buying new equipment based on a single result.
An internet speed test follows the route used by this browser, including a VPN or proxy when applicable. Compare enabled and disabled states only if changing that setting is appropriate for your situation. Record both states clearly. The test itself does not change your VPN, choose a tunnel location or inspect traffic from other applications.
A faster internet speed test result with a VPN can occur when routing changes; a slower one can reflect added distance, congestion or processing. Neither result establishes whether your privacy protection is working. Use IP comparison to examine destination-specific public addresses and the WebRTC check for the separate evidence that it collects.
An internet speed test measures a path between endpoints, not every path on the internet. A nearby edge, a distant file server and a heavily used service can perform differently. Compare the named sources and methods before treating a difference as an error. Multiple providers can add useful evidence when they are tested under similar conditions.
The internet speed test cannot confirm your access technology, subscription speed or precise measurement-server location from throughput alone. Cloudflare routes traffic within its own network. We display the endpoint actually requested and do not invent a city, distance or dedicated server identity from your current public address.
This internet speed test does not send ICMP echo packets or run a UDP packet-loss experiment. Failed HTTP requests are not converted into a packet-loss percentage. A game may use another destination and protocol, and its displayed latency may reflect additional processing. Treat the HTTP measurements as one useful observation rather than a substitute for the game’s own diagnostics.
For route investigation, pair the internet speed test with Ping or traceroute. Those tools originate from a named remote probe, so their direction and source differ from this browser-based test. Reading the source labels prevents an apparent contradiction when two tools examine different network paths.
Starting the internet speed test reveals your public address and ordinary request information to the measurement provider. MyIP.dog sends generated test payloads, not selected personal files, and does not submit a separate measurement report to Cloudflare’s results-logging endpoint. Cloudflare still processes the requests under its own policies; this is not an anonymous test.
The internet speed test keeps its displayed snapshot in page memory. Copy report copies the current result and methodology to your clipboard only when requested. The report contains no visitor IP field, but its time and performance may still reveal context when shared. Leaving the page clears the displayed snapshot, and returning does not silently restart a measurement.
Save an internet speed test report with a short note about location, connection type, VPN state and background activity. Change one factor at a time. If a problem consistently appears at certain hours or in a specific room, that pattern is more actionable than chasing the largest number observed during repeated tests.
An internet speed test is most useful when it leads to a specific next check: improving Wi-Fi placement, pausing an upload, examining a VPN route or gathering evidence for your provider. Keep the raw samples and the measurement scope with your conclusions. Repeatable observations make a stronger case than a single unexplained score.
This internet speed test uses Cloudflare’s public download and upload measurement endpoints, with MyIP.dog’s own bounded browser transport and calculation method. Its results are not produced by the Cloudflare Speedtest SDK and are not directly interchangeable with that product’s default test. Review Cloudflare’s measurement-service description, the official endpoint configuration, and Resource Timing documentation. Our methodology and data sources explain how this tool fits the rest of the site.