The “HTTP Archive” format

Hi Pat,

How can we download a waterfall directly using WPT ? Right now there is an option - “Export HTTP Archive (.har)” which is very much like what’s available in chrome but the problem is its quite difficult to find an executable which opens this file format. Can we somehow download the waterfall Time graph in common format like jpg because I just don’t know whether it’s directly accessible via WPT?


There is a HAR viewer - but generally we need some better tools for analysing and comparing HAR files

Can we somehow parse the HAR files and plot graphs of individual element response times ?

For instance if we want to see a trend of response times for each element and graph them individually, can this be achieved by parsing the HAR files ?

Not sure exactly what you are looking for. The png’s for the waterfalls are readily available (the API also includes links for them) though it won’t plot anything at an individual element level over time.

If you want to watch specific elements and plot them over time you would need to parse either the XML from Webpagetest or the HAR’s and implement the plotting logic yourself.

You’ve got that right Pat. I would like to download data in XML format so that i can write a logic to plot the performance of each element over a period of time. The reason for this is, i would like to monitor the trend of each element and compare which element takes more time to load and hence ending up degrading the overall performance of the page (talking shear from the page load angle).
I’m not much familiar with HAR file format and i do not know how to parse it but i’m quite comfortable with XML.
Can you tell me where exactly can i get data in xml format using WPT?
And if its possible to implement what i’m thinking about??

Here are the docs (for what they are) on the API which includes the XML interface:

Perfect! I’ll go through these and revert if i have any more question. Thanks a ton for your help with this Pat!

The API converts the result summary page into XML tags but i don’t see it translating the waterfall into xml.

For example - The page can be converted into xml like this:

but this page
cannot be converted in the similar way -

I’m more concerned in converting the waterfall result into xml where i can get the load times of individual elements.

Sorry, try throwing a &requests=1 on the end of the XML request - that will get it to include the raw request data in the XML as well (not enabled by default because it’s a fair bit slower). I’ll make sure to add that to the docs:

Docs updated:

There are a few other parameters that also add more data to the XML responses.

It might be worth taking a look at HAR Storage ( and seeing whether that will help you

Thanks Patrick! This really Helped!!