Best 3 HAR File Viewers Compared

A user says the page is slow. You cannot reproduce it, so you ask them to export a HAR file from their browser — and now you have a 40MB JSON document describing four hundred requests.

Before opening it anywhere: a HAR capture contains everything the browser sent and received, including cookies, session tokens and anything posted in a form. It is a credential dump wearing a debugging tool’s clothes. Where you open it matters more than which features the viewer has.

How to judge a HAR viewer

Does the file stay local? The single most important question. A HAR from a logged-in user contains live session cookies.

Is the waterfall readable? Sequence and overlap are the point. A sortable table tells you sizes; only a waterfall shows you what was waiting for what.

Does it break down timing? DNS, connect, TLS, TTFB and download are different problems with different fixes. A single “duration” number hides which one you have.

Can you filter? Four hundred requests is unmanageable without filtering by status, type or domain.

The comparison

Facts checked August 2026; tools change. Table covers only the 3 alternatives — our tool gets its own section below.

Unlighthouse HAR Viewer

The most detailed on timing, which is where the answer usually is. Its waterfall visualises the sequence and overlap of requests, and it breaks each one into blocked, DNS lookup, connect, SSL/TLS, send, waiting (TTFB) and receive phases — the distinction that separates “the server is slow” from “we opened six connections to a third-party domain first”.

It also breaks down resources by domain, runs entirely in your browser with the HAR never uploaded, and is free. Notably, its own page warns that HAR files capture full request payloads including plaintext passwords in postData and active session cookies — a warning more tools should carry.

HAR Analyzer

The best at cutting a big capture down. Filtering covers URL search, status code, HTTP method, content type and domain, and you can inspect request details, headers, response previews and timings per entry. For a capture with hundreds of requests, filtering to “4xx and 5xx on our own domain” is the fastest route to the actual problem.

Processing happens 100% in your browser memory, with the file analysed locally and never leaving the device, and there is no account. File size is bounded by your device’s RAM rather than an arbitrary cap — generous, though a very large HAR on a modest laptop will still struggle.

SpeedVitals

The quickest read. It turns a capture into an interactive waterfall covering documents, scripts, styles, images, fonts, API calls and third-party resources, with request timing, status codes, transfer sizes and connection phases.

Its framing is page performance rather than debugging, which shapes what it surfaces — third-party resource weight, for instance, is prominent. It parses the HAR locally in your browser so the file is not uploaded, explicitly to protect cookies and tokens. No account mentioned. Free.

HAR File Viewer

Ours opens a HAR capture and reads requests by timing, size and status, with the waterfall and the slow entries surfaced first. Surfacing the slow entries by default is the deliberate difference — you did not open the file to browse four hundred requests, you opened it to find the two that mattered, and sorting for that is one fewer step.

What it does not do: break timing into every individual phase the way Unlighthouse does, or offer filtering as extensive as HAR Analyzer’s. For a deep timing investigation those go further. Everything runs in your browser, which for this file type is not a nicety — a HAR file is a complete record of a browsing session, and uploading someone else’s to a third party is a genuine disclosure.

Which one to pick

• Working out which timing phase is slow — Unlighthouse’s breakdown.
• A huge capture that needs narrowing — HAR Analyzer’s filters.
• A fast look at page weight and third parties — SpeedVitals.
• Finding the slowest requests immediately — the tool below.

How to do it with HAR File Viewer

1. Open the HAR File Viewer and load the capture — it stays in your browser.
2. Start with the slowest entries rather than the top of the list.
3. Check whether slow requests are waiting on the server (TTFB) or on the connection before it.
4. Look for requests that only started after something else finished; that is a dependency chain, and it is usually fixable.

The walkthrough is in how to read a HAR file and find slow requests. Other developer tools are in the tools directory.

You might also need

The HTTP Headers Parser helps make sense of the caching and security headers a HAR entry shows.

If the investigation turns out to be about HTTPS behaviour rather than speed, the HTTPS Checker covers redirects, HSTS and mixed content.

Frequently asked questions

Is there a free HAR viewer that doesn’t upload my file?

Yes. Unlighthouse, HAR Analyzer, SpeedVitals and ours all state the file is parsed in your browser. Given what a HAR contains, treat that as a requirement rather than a preference.

Are HAR files sensitive?

Very. They record request and response headers including cookies and authorization tokens, and often POST bodies containing credentials. Never post one in a public issue tracker, and scrub it before sharing privately.

What should I look at first in a HAR file?

The slowest few requests, then whether their time is spent waiting for the server or before the connection is established. High TTFB points at the backend; long blocked or connect phases point at too many origins or connection limits.

Final thought

Ask where the time went before asking what to optimise. A capture full of fast requests that ran one after another is a completely different problem from three slow ones running in parallel, and the waterfall shows you which you have in about ten seconds.

Try the free HAR File Viewer