Analyze a browser recording
Export a HAR from your browser or record a full HAR in Playwright. Load it below, then choose one byte field for the entire analysis. The optional comparison file should record the same task under the same test conditions.
Understand the two byte measurements
Response body bytes sums nonnegative response.bodySize values. Browser transfer bytes sums the exporter-specific response._transferSize field. The analyzer never substitutes decoded content.size for a missing transfer measurement.
Missing values and -1 remain unknown. A recorded zero is retained as zero; it does not establish why no bytes were recorded. The percentage comparison is available only when both recordings have complete coverage of the chosen field and the baseline is greater than zero.
These are recorded response measurements. They do not reconcile request uploads, proxy tunnels, retries outside the recording or your provider’s billing meter. Use account usage data for a billed-traffic estimate in the proxy bandwidth calculator.
Capture a comparison you can use
- Choose an observable pass condition, such as the expected product count or a required text value.
- Run a baseline with the same browser version, viewport, cache policy and service-worker settings you will use in the comparison.
- Change one resource-blocking rule, repeat the task and check that the required output still exists.
- Compare both recorded totals and unsuccessful responses. A smaller file can simply mean the task failed earlier.
The Playwright bandwidth guide includes a bounded script and explains why enabling request routing changes cache behavior. For HAR recording, Playwright’s full mode retains sizes that minimal mode omits.
Keep the recording under your control
Chrome offers a sanitized HAR export that removes some sensitive headers. Review the file before sharing it: URLs, query strings or response content can still contain private information.
This analyzer reads the file in your browser. Its results use domain names and counts; they omit URL paths, headers and bodies. The downloadable summary contains domain names, which may also be sensitive. The clear button removes the tool’s current inputs and results.
After measuring a valid task, move from the traffic estimate to current mobile proxy prices and stock.
Common questions
Does the analyzer connect to my proxy?
No. It reads the HAR you select. Recording the traffic happens in your browser or application before you use this page.
Why is the reduction percentage unavailable?
The selected byte field is missing from some requests, or the baseline total is zero. Check the measured-request count in each recording before comparing totals.
Can HTTP 200 still be a failed task?
Yes. Status alone does not show whether a response contains the expected result. Validate the application output separately.
Sources and review scope
The calculation uses the selected fields in your file. The example is a fictional fixture, not a result from the Proxies.sx network. Official exporter and browser references were checked on 13 October 2026.
- Chrome DevTools HAR export
- Chrome DevTools HAR exporter implementation
- Playwright HAR recording
- Playwright request size fields
Agent service index and Pool Gateway documentation provide the API entry points. The tool index as JSON lists these tools and their limits.
