QRX doesn't need HTML body

It all started when I saw this error in the df console:

622|direct | [DISCOVER HTML FAILED] [http://thesecularfoxhole.live/](http://thesecularfoxhole.live/) Request failed with status code 503 (Service Unavailable): GET [https://www.thesecularfoxhole.live/](https://www.thesecularfoxhole.live/)
622|direct | [DISCOVER HTML FAILED] [https://www.thesecularfoxhole.live/](https://www.thesecularfoxhole.live/) Request failed with status code 503 (Service Unavailable): GET [https://www.thesecularfoxhole.live/](https://www.thesecularfoxhole.live/)

The strange part was that the website was fully accessible in the browser, and the auto-discovery tag was also there:

<link rel="alternate"
      type="application/rss+xml"
      title="The Secular Foxhole"
      href="https://feeds.captivate.fm/the-secular-foxhole/" />

At that moment it became clear that the problem was on our side. There were two possible explanations, and increasing the timeout immediately solved the issue. Previously it was 3 seconds, and the site simply did not have enough time to load completely. And the word "completely" is the key point here.

To extract flows from <head>, QRX does not actually need to download the entire HTML page. We only need the document until </head> or </body>.

And a check in PageSpeed Insights showed that thesecularfoxhole.live had problems specifically in "Performance".

thesecularfoxhole.live PageSpeed Insights

The solution

Now QRX no longer waits for the page to fully load. (It would actually be interesting to understand why it waited in the first place. Why did we need the full page?)

As soon as we receive something like:

const HTML = [
  "</head>",
  "<body"
];

or if the entire page becomes too large while there is still no <head> ending, we stop reading. There is nothing more for QRX to discover at that point. Full HTML loading may appear later in QRX, but for now the <head> section is enough.

Meet QRX 0.8.1 "No body".

Published on May 27, 2026

qrx