Methodology & data sources
Our reports combine clearly identified public datasets and real-world measurements. Each metric is presented with its scope, freshness and limitations.
Browser connection check
The homepage, fiber check and multi-gigabit check transfer data over HTTP between your browser and this site. Download and upload measure transferred bytes divided by elapsed time, including request overhead. Each direction runs for about five seconds or until 256 MiB has been transferred. A slow request can extend the test; the overall timeout is 60 seconds.
Response timing is the average elapsed time of successful HTTP requests, including server processing. Response variation is the mean absolute difference between successive successful samples. Failed HTTP requests are counted separately and are not network packet loss. At least two valid response samples are required. The check does not measure loaded latency or identify a faulty provider from a single result.
The result can be saved for a signed sharing link. Shared summaries and downloads omit IP and location; the server processes connection information as described in the privacy policy. In local development, this checks the Docker endpoint and cannot measure an internet service plan.
M-Lab speed test
The dedicated speed test and embeddable widget use Measurement Lab NDT7 for throughput. M-Lab uses its own test infrastructure and data policy. Compare measurements made with the same method and similar conditions; an HTTP check to this site is not an NDT7 result.
Broadband availability
United States provider, technology and coverage information comes from public FCC Broadband Data Collection releases. Provider-reported availability describes where service is offered; it does not guarantee installation at a specific address.
Measured performance
Real-world speed and latency summaries use measurement datasets such as M-Lab where available. Measured results and advertised speeds are kept separate because they answer different questions.
Aggregation and freshness
Location-level public records are aggregated to appropriate geographic levels such as state, county or city. Pages identify the reporting vintage and our import date so readers can judge freshness.
Corrections
If you identify a mismatch, contact us with the page URL and supporting evidence. We review material corrections and update traceable source information when appropriate.