HAR File Analyzer

A HAR file is a JSON-formatted record of network activity captured by a browser or debugging tool. It can contain request URLs, methods, HTTP status codes, response headers, MIME types, transfer sizes, initiators, redirects, and timing phases for every resource loaded during the recording.

The analyzer converts this raw data into an interactive network waterfall. Requests are aligned on a shared timeline so you can see what started first, what waited, which resources overlapped, and which late request extended the page load. This makes a complex capture easier to review outside the original browser session.

Sort and inspect entries to locate long server waits, slow downloads, large JavaScript bundles, unoptimized images, late-loading fonts, duplicate requests, redirects, and client or server errors. Status codes such as 301, 404, 429, and 500 often reveal broken dependencies or unnecessary work that a simple visual page review misses.

Group patterns by hostname and resource type to understand the cost of analytics, advertising, tag managers, customer support widgets, and other third parties. A single request may look harmless, while its redirects and dependent downloads create a much longer critical request chain.

A waterfall can separate queueing, DNS lookup, connection setup, TLS negotiation, request sending, server wait, and response download. A long wait phase often points to server processing, origin latency, or an overloaded third-party service. Repeated DNS, connection, and TLS phases can indicate weak connection reuse or requests spread across too many hostnames.

Timing fields depend on what the capturing browser exposes, so not every HAR includes every phase. Use the available values as evidence, then confirm important findings with a fresh capture and server, CDN, or real-user monitoring data.

The position of a request in the waterfall matters as much as its duration. A hero image that starts late may be hidden behind a stylesheet, script, client-side render, or CSS background declaration. A font may be requested only after CSS arrives. Redirects add serial round trips before the final resource can begin.

Use the waterfall to form a hypothesis, then run the LCP Finder to identify the Largest Contentful Paint element and its timing breakdown. Combining element evidence with request timing makes it easier to prioritize preload, fetch priority, compression, caching, or application changes.

Open the browser developer tools, select the Network panel, preserve the log if the journey spans navigation, and start recording before reproducing the issue. Reload the page or complete the exact interaction, wait for the relevant requests to finish, then export the capture as HAR. Record whether the browser cache was enabled and whether the session was authenticated.

One capture represents one environment and one journey. Compare cold and warm cache visits, mobile and desktop conditions, or working and failing sessions when those differences matter. This is useful for page speed audits, crawl troubleshooting, API failures, checkout errors, and user-reported loading problems.

SpeedVitals parses the selected HAR file locally in your browser, so the file is not uploaded for analysis. This matters because HAR files may contain cookies, authorization headers, session tokens, query parameters, form data, internal hostnames, and personal information.

Local analysis does not make a HAR safe to share. Remove or redact credentials and private values before sending a file to another person or attaching it to a ticket. If a capture contains an active secret, revoke or rotate it according to your organization's security process.