Ask HN: Is there a browser plugin which replaces webpage with its screenshot? | Hacker News Reader

Ask HN: Is there a browser plugin which replaces webpage with its screenshot?

4 pointseimrine5 comments

I am being exhausted from observing interactivities of some webpages which are typically being fired while scrolling the webpage. Typically the interactivities are some pesky features like demonstrating me a video or some banner which wants some actions from me, typically expecting me to click "X" to get rid of the pesky demonstration. All I want from webpages is plain text, but I can agree with rich text - different fonts, with colors, with shadows, with stupid pseudoelements, you know what I mean. I have discovered some relief from reading links from HN as screenshots and all I want is a little bit optimization of obtaining screenshot instead of the webpage in a browser. Any plugins for any browser capable to do this?

View on HN

Not exactly screenshot, but for Firefox on Windows(10Pro) I have installed the "Open in Reader View" add-on. So when I see an article I want to open in a new tab, I right click on the link, on the menu that appears I choose the "Open in Reader View" and it opens a new tab with the article/page in Reader View. It is the cleanest way I access pages.

Before Firefox added the feature there were a couple of browser extensions called 'readability'. This for example can enable it per domain, there might be some who can do it for every website https://chromewebstore.google.com/detail/permanent-readabili...

This is a fascinating idea.

I think you could hack a way to do this with BrowserBox: which, funnily enough, already "replaces web pages with their screenshots" (we stream the screenshots or screencast from the headless browser and serve it).

However to get what you really need, I see some difficulties:

- you could scroll the entire page first, stitch the screenshots together and then not update any further, but this breaks in some weird edge cases:

  - dynamically loaded content which changes the scroll height as you move down, will not be loaded or displayed in this method reliably. 
  - this height change will then throw off any clicks that occur below it
  - this will still not prevent banners and the like from occurring

- you could use the --print-to-pdf command line flag for Chrom/Edge/Brav/ium browsers. Not sure if hyperlinks work from these. But you could set up a basic proxy to take all HTTP and HTTPs requests and pass them through your headless "to pdf" render process if they do.

- if hyperlinks don't work in these pdfs you could use Playwright (or CDP if you're braver), to evaluate "document.links" property (see: https://developer.mozilla.org/en-US/docs/Web/API/Document/li...), and just create a simple "list of links" HTML page templated from these and append that to the bottom of your PDF document (by using some command line tool in collab with your to-pdf proxy).

Anyway, fascinating idea, good luck with it! :)

It's not exactly what you're asking for, but a means to a similar idea... I setup a "site search" in Chrome for archive.is so that I can just type `a ` before the URL in the location bar and have it auto-complete the archive.is URL for a page.

    Search engine: archive.is
    Shortcut: a
    URL with %s in place of query: https://archive.is/?run=1&url=%s&

This works to effectively make 99% of pages static.

Based on past experience, I believe the archiver allows JS during load but is disabling JS during runtime (but I haven't actually verified that).

There is also an unofficial Chrome extension with similar functionality, but it is closed source and requires entire browser history permission for some reason, so...

Other than reader mode, you can try disabling javascript. Lynx will get you plain text, but usability will vary from site to site.

If you truly want a screenshot, you can write a program to spin up a headless browser / selenium, load the page, and save the result somewhere. People do this all the time as part of UI and integration tests and tooling is decent enough.

Reply on news.ycombinator.com