hoakley July 1, 2022 Macs, Technology

How should you save web pages in Safari?

The Internet is frighteningly ephemeral. The site you found so useful today may have vanished next week, and the chances of it still being there in a couple of years may be slim. This article looks at how best you can keep a more permanent record of individual pages, rather than whole sites, including longer-term archiving.

Other browsers offer different options, but in current releases of Safari your choice for saving web pages includes:

File/Save As…/Page Source to save the page as a single HTML source file.
File/Save As…/Webarchive to save the page as a single Webarchive file.
File/Export As PDF… to save the page as a single PDF file, in display format.
File/Print…/Save as PDF to save the page as a single PDF file, in print format.

HTML source

This is the smallest and least complete version of the four, as it contains the HTML source of the page, and omits all linked and similar generated content. For relatively plain pages containing text exclusively, this can be useful, but that is now appropriate for few pages. The saved file can be opened in Safari or another browser, and so long as none of the linked content is missing or changed, you should see the original page reconstituted faithfully.

This is almost certainly not what you’d want as a more permanent record such as an archive.

Webarchive

This is an opaque file format that assembles the entire content of the page, including embedded images and other content, but not linked downloadable files, in a single incomprehensible text file.

Although this format is peculiar to Safari, it has limited support by some other browsers, which can read it, and a few utilities which can, for instance, convert a Webarchive to PDF. However, with a single exception, those utilities are now old and their future uncertain. It’s probably best to consider this format as being proprietary to Safari.

I’ve not been able to find any description of the format which could be used to implement a third-party reader, and as it doesn’t appear to comply with any open standard, the format may well have changed over the last few years. In the past it has been well-supported by the macOS API, but currently all those existing calls to work with Webarchive files are marked as being deprecated by Apple, so are likely to be removed whenever it wishes, making it impossible to use those deprecated calls in future versions of macOS.

The good news, though, is that Webarchives are now making their way into Apple’s open source WebKit. Currently, support is limited to writing but not reading them from WKWebView from macOS 11 onwards, but hopefully that will be extended in the future.

Because Webarchives contain all linked content, and are encoded in text representation of binary data, they’re by far the largest of the four options. They’re probably the best way to save web pages for the time being, but their format may well not be supported in just a few years, so they’re unsuitable for archives intended to last more than that period.

PDF

In case you weren’t aware, there are two different routes in Safari to turn a webpage into a PDF document: directly using the Export As PDF… menu command, and indirectly via the Print… command then saving as PDF from the Print dialog. The results are quite different.

safaripdf1

Exporting as PDF creates a document in which the entire web page is on a single PDF page, which may not be what you had in mind. The advantage of this is that the PDF is one continuous page without any breaks, and is a faithful representation of what you see in your browser, complete with its original layout and frames. The disadvantage is that this won’t print at all well, imposing page breaks in the most awkward of places. Very long pages can also prove ungainly, and difficult to manipulate in PDF utilities.

safaripdf2

Printing to PDF breaks the web page up into printable pages, and splits up frames. What you end up with isn’t what you see online, but could at a push be reassembled into something close to the original. That isn’t too bad when the placement of frames isn’t important to their reading, but if two adjacent columns need to appear next to one another, this layout is likely to disappoint. It is the best, though, for printing, with headers and footers and page numbering too.

Recently some have remarked that PDFs generated from Safari lose their embedded links. That shouldn’t happen: well-constructed web pages should preserve all their original links when viewed as PDFs.

PDF documents are easy to create in macOS, as Safari uses Quartz graphics to render web pages, and PDFKit is also part of Quartz graphics. The two alternatives result from Safari rendering its displayed content through PDFKit, or rendering its content as prepared for printing.

While PDF is one of the preferred formats for archiving laid-out documents, it’s worth bearing in mind that standard macOS PDF isn’t compliant with any of the PDF/A standards for archival documents. You’d need a high-end PDF editor such as Adobe’s Acrobat (Pro) CC to prepare and save to any of those.

Conclusions

Save As…/Page Source is of limited use, mainly for text-only pages without embedded content such as images.
Save As…/Webarchive is excellent for day-to-day use, being complete and faithful, but isn’t an open standard and its future may be limited. It’s therefore not recommended for archival purposes.
Export As PDF… is excellent for day-to-day use, complete and faithful, but for serious archival use needs to be converted to comply with an archival standard in the PDF/A series.
Print…/Save as PDF is an alternative most suitable if you’re going to want to print the document out.

38Comments

Add yours

1

Paul D. on July 1, 2022 at 6:52 am

A few comments on things that may or may not be obvious.

1. Some websites use an @media rule that makes the page display differently when printed, so Export as PDF and Print / Save as PDF might produce very different-looking results.

2. Two related options are Export as PDF and Print / Save as PDF from the Reader View, which is often the best choice for searchable, easy-to-read PDFs with small file sizes.

LikeLiked by 1 person
- 2
  
  hoakley on July 1, 2022 at 7:02 am
  
  Thank you.
  Howard
  
  LikeLike
3

EcleX on July 1, 2022 at 6:58 am

Thanks for the useful information. You said “Webarchive. This is an opaque file format that assembles the entire content of the page, including embedded images and other content, but not linked downloadable files, in a single incomprehensible text file”.

In my experience, sometimes linked downloadable files (or links in general) do not work as you said, but other times they do. In other words, sometimes such links point to the original source (and work), but other times they point to where you saved the “.webarchive” file in your disk (which can be seen hovering the arrowhead mouse pointer over such links in the saved “.webarchive” file once opened by Safari). I think that the issue stems from the way the original web pages were built.

LikeLiked by 1 person
- 4
  
  hoakley on July 1, 2022 at 7:04 am
  
  Thank you. That’s a bug, I think, although whether that’s in the page or Safari isn’t easy to work out.
  Howard
  
  LikeLike
5

Aaron Shepard on July 1, 2022 at 7:23 am

My understanding is that the chief characteristic of PDF/A is that it does NOT embed any of the 13 most common computer typefaces — that is, the thirteen most common at the time the standard was developed. This was not to enhance its value as an archive but as a means to slightly reduce file size — a negligible benefit nowadays. PDFs by Apple, by contrast, seem to embed ALL fonts, which makes them SUPERIOR as an archiving format. About the only use for PDF/A nowadays seems to be to fool people into thinking they’ve embedded all fonts when they haven’t. It’s a very common myth among self publishers, for example. Trying to use such files for commercial printing, they can run into all kinds of trouble. If there’s a convincing reason to still use PDF/A for anything at all, I’d honestly love to hear it.

LikeLiked by 1 person
- 6
  
  hoakley on July 1, 2022 at 7:36 am
  
  Thank you.
  Well, my understanding, and that detailed in Wikipedia’s account, is the exact opposite: all fonts used must be embedded, as the goal is to produce a standalone file which doesn’t rely on external font resources of any kind.
  Whether you like any of the four variants of PDF/A or not, they’re established ISO standards which are widely used for archival purposes. In fifty or a hundred years time, they’re the most likely to still be fully accessible and supported, which is their purpose. They’re not intended for print or pre-print, so only the seriously misguided would try using them for that purpose. As I’m sure you’re well aware, there are other PDF standards in PDF/X which are intended for that, but not for archiving.
  Using the right tools and standards for the purpose is very important when you’re looking into the more distant future, assuming humans ever get there.
  Howard.
  
  LikeLike
  - 7
    
    Aaron Shepard on July 2, 2022 at 12:32 am
    
    This is exactly the disagreement I had recently in a Facebook author group, which ended with the author creating a test file and discovering for himself that I was correct. The omission of the most common fonts — like Times New Roman and the others that Microsoft made profilic — is not a secret. It’s the way the format was explicitly described by Microsoft when they introduced saving as PDF/A into Word. Anyway, you may wish to test this for yourself, because if you’re using this for archiving, you may not be getting what you think.
    
    LikeLike
  - 8
    
    Aaron Shepard on July 2, 2022 at 1:11 am
    
    By the way, my comment about authors using PDF/A was not meant to say they SHOULD be able to use it — obviously, that’s a mistake from the beginning — but that PDF/A files often were rejected or caused characters to go missing specifically because not all fonts were embedded — exactly opposite to what the authors expected and the reason they were told the format would work for them. Often this showed up in characters like list bullets or numbering, which Word renders by default in a font not chosen by the author. One reason that PDF/X is recommended for printing is that, by contrast with PDF/A, it DOES embed all fonts, with no exceptions, because the files are not meant just for viewing on the desktop. Unfortunately, most authors do not have Acrobat Pro for preflighting, so they never checked the embedding in their PDF/A files.
    
    LikeLike
  - 9
    
    Aaron Shepard on July 2, 2022 at 4:15 am
    
    Well, Howard, I learned something new. It turns out we’re both right. You’re right that the standard requires all fonts to be embedded. However, “not all PDF/A creation products respect the requirement to embed the 14 Adobe base fonts.” (See https://www.soliddocuments.com/iso-19005-1-compliance.htm?subject=6.3.4). So, yes, PDF/A should have all fonts embedded, but that’s not what you always get. You would have to test your specific app for compliance. And Microsoft apparently fails to comply, or at least did for some time. (Big surprise.)
    
    LikeLiked by 1 person
    - 10
      
      hoakley on July 2, 2022 at 5:19 am
      
      Thank you.
      I wrote unambiguously: “While PDF is one of the preferred formats for archiving laid-out documents, it’s worth bearing in mind that standard macOS PDF isn’t compliant with any of the PDF/A standards for archival documents. You’d need a high-end PDF editor such as Adobe’s Acrobat (Pro) CC to prepare and save to any of those.”
      The word “Microsoft” doesn’t appear anywhere in this article. This article is about Safari, which isn’t and never has been a Microsoft product. Microsoft doesn’t offer any product which could in the wildest stretch of the imagination be described as “a high-end PDF editor”.
      Howard.
      
      LikeLike
11

George Mirabelli on July 1, 2022 at 8:06 am

If you open the developer console, you can choose the exact node you want to print, right click and choose print to PDF, this allows you to print the exact frame you want.

LikeLiked by 1 person
- 12
  
  hoakley on July 1, 2022 at 8:16 am
  
  Thank you.
  Howard
  
  LikeLike
13

rfog926695139 on July 1, 2022 at 9:14 am

I use reader view and then send to DEVONthink via context menu. I did a keyboard shortcut (CMD+D) and directly sent to DT that way.

In combination with DT, you could capture the web in “Formatted Note”, that is a all-in-one embedded HTML page compatible with all Web Navigators (even under Windows). It is a completely self-contained file similar to Webarchive but completely standard.

LikeLiked by 1 person
14

Alan Ralph on July 1, 2022 at 9:23 am

I’d suspected that the WebArchive format might be an Apple-only thing, thanks for the confirmation.

These days I use the Web Clipper that comes with DEVONthink to grab web pages that I may want to refer back to later, partly because it’ll work regardless of which browser I’m using (currently Firefox), and partly because it supports a range of capturing options, not just WebArchive but also Markdown text and PDF document. For some of those options there is also the ability to produce a ‘clutter-free’ version, similar to what you’d get using Reader view.

Usually this works without issue, but I’ve found it prudent to hit the Preview button in the Web Clipper to check what is being captured (not all capture formats support this option, alas.) Depending on the site and page design, information may be omitted or not present at all. Sometimes turning off ‘clutter-free’ resolves this, but in some cases I’ve had to resort to saving pages as PDF files from the Print dialog.

Most of the time I use WebArchive, but that is only used as an intermediate step — anything I want to keep for reference gets reviewed, tidied up if necessary, then converted to PDF in DEVONthink.

I find it deeply ironic that the ‘old web’ presents the least resistance to saving information for archival. Alas, a lot of ‘modern’ web site creation seems to produce pages that lack the semantic markup required to capture the pertinent information, or worse still refuse to display at all without JavaScript or cookies.

LikeLiked by 1 person
15

Tom72 on July 1, 2022 at 12:47 pm

When I’m developing a talk I’m going to give, I’ll try to use Reader View and the PDF link from the print dialog in Safari. As Alan Ralph said it produces a “clutter-free” output. It also has the link to the page at the bottom for future reference. Sometimes the output pdf has missing bits the page breaks. In these cases I use Safari’s Paper Size option in the print dialog to make longer page lengths. I have a saved set of several LetterX2 (8.5X22), LetterX4 (8.5X44), etc. These enable making a set of pdf’s I can use to copy a bits from the web page and I have the reference to use in my talks. It also doesn’t require any other software.

LikeLiked by 1 person
16

Milo on July 1, 2022 at 1:54 pm

I can recommend the “Single File” Extension for Firefox and Chrome. It produces a single file with media encoded inside the HTML for portability.

https://github.com/gildas-lormeau/SingleFile

There is a comparison of different archiving file formats in the README as well.

https://github.com/gildas-lormeau/SingleFile#file-format-comparison

Unfortunately it does not support Safari at this time. Likely because of Apple’s restrictive distribution policy for WebExtensions. Safari is able to open the files produced by Single File, as it’s standard compliant HTML and CSS.

LikeLiked by 1 person
- 17
  
  NSSynapse (@NSSynapse) on July 1, 2022 at 8:50 pm
  
  I wholeheartedly second this recommendation. I’ve been using the SingleFile extension with Vivaldi for some time now, and I love how it creates neat single files that perfectly preserve the page layout, preserve text as text (instead of just being an image), all while using a standard format that every browser can render.
  
  LikeLiked by 2 people
  - 18
    
    hoakley on July 1, 2022 at 9:21 pm
    
    Thank you, but the title of this article establishes that this is about Safari, which remains the most popular browser on Apple’s products.
    I’m also not sure what you’re referring to when you write “preserve text as text (instead of just being an image)”. Both Webarchive and produced PDF do exactly that, and preserve page layout too. I don’t know of any direct way of rendering Safari’s web pages into text as images, do you? The reason for that is Quartz, used by Safari to render pages on the display, which handles text as text and not as images.
    Howard.
    
    LikeLike
    - 19
      
      NSSynapse (@NSSynapse) on July 1, 2022 at 10:08 pm
      
      I’m sorry, I didn’t mean to distract from the article itself and the valuable analyses you have done regarding the various options that Safari provides in this area.
      
      The HTML files that SingleFile creates are, of course, fully compatible with Safari, and I personally only save a web page every week or two, so briefly switching to a different browser to save a web page would not be much of an inconvenience. So my thinking was that if I used Safari instead of Vivaldi or if SingleFile was only available for Safari, I would still use the extension to save web pages, even though I would have to briefly switch browsers.
      
      This is largely driven by my personal preference to save web pages as HTML and not PDF, though, and I might have gotten carried away a bit by my enthusiasm for how well the SingleFile extension matches my personal needs and preferences.
      
      The “preserve text as text (instead of just being an image)” part was not referring to any of Safari’s options, but to saving a full web page as an image file (i.e., basically as a screenshot of the whole page). Vivaldi has a built-in function to do this and there are extensions for other browsers to do the same. This gives you a pixel-perfect representation of what you saw, but having the web page’s text stored as pixels in a bitmap image instead of as actual text is, of course, a pretty big downside.
      
      LikeLiked by 2 people
    - 20
      
      hoakley on July 2, 2022 at 5:13 am
      
      Thank you. No problems, and no need to apologise.
      It’s worth remembering that Safari uses Quartz to render pages to both the display and to PDF files, so the results should be essentially identical. Other browsers might work differently.
      Howard.
      
      LikeLike
21

Dakotamoons on July 1, 2022 at 2:18 pm

Also good is Evernote’s Web Clipper extension which saves various selections, including links.
My primary browser is Brave, but I have the EN extension installed in Safari as well.
Seems to work reliably.

LikeLiked by 1 person
22

Toby Sanchez on July 1, 2022 at 2:39 pm

The good news about Webarchives is that most applications that preview any document will preview webarchives, on Macs of course. I archive web pages in webarchive format and store them in DevonThink for later processing. DevonThink allows EDITING of webarchives in a WYSIWYG fashion, after which you can keep as a webarchive or convert to HTML or to PIXEL PERFECT PDFs. I highly recommend DevonThink for manipulating webarchives. PathFinder is useful for viewing them. Some Linux webkit based browsers will open archives ( I think MIdori does ) and/or Safari for Windows can be run under WINE to view webarchives directly in Linux.

LikeLiked by 1 person
- 23
  
  hoakley on July 1, 2022 at 4:20 pm
  
  Thank you.
  Yes, as I wrote, other apps can access Webarchives. But that doesn’t make them suitable for archival use, to access the content in 5-50 years time. Apple could help this by making the format an open standard, which would allow anyone in the future to access them. Sadly, of the formats which Safari can currently write to, only PDF has that property, and even then needs to be made conformant with one of the PDF/A standards to be reliable.
  Howard.
  
  LikeLike
24

Duncan on July 1, 2022 at 2:51 pm

Howard, thank you (personally) for pursuing this. Unfortunately I’m really busy this morning so I can’t keep up with all the comments, but I’d like to ask about this:

“Because Webarchives contain all linked content…”

I’m guessing this means a webarchive could present the original page, nearly complete, if opened on a Mac that is not connected to the Internet?

One other benefit of a webarchive is it will show the page in the Finder’s QuickLook (space-bar) feature.

My inclination is to use Webarchives to save important pages (despite their size), with the hopes that there will be a utility in the future that can convert them to something more open. And even if a future version of Safari can’t open them, as a last resort someone could set up a virtual machine with an older OS that contains a copy of Safari that can. Better than nothing.

LikeLiked by 1 person
- 25
  
  hoakley on July 1, 2022 at 2:56 pm
  
  Yes – that’s the purpose of the format, to be as self-contained as possible.
  It would be really helpful if Apple were to make this an open standard.
  Howard
  
  LikeLike
26

Joey Jay on July 2, 2022 at 2:06 pm

Thank you yet again for seeding new articles which appear to be organically based on issues raised in the comments section of others, such as the one on archiving.

I am in the market for a new roof on my house and the salesperson was on his way. I quickly needed to print out an entire thread of conversation about metal roofs from the NextDoor website, a social media service for neighborhoods to discuss various topics of concern to residents in the USA.

In iOS, nothing captured the entire thing properly. A quick search revealed a 3rd party app called Instaweb, which prints web pages after rendering them in its own built-in browser. It got the job done quickly and flawlessly so I thought I would mention it as an option. There are other apps, but this one has good feedback ratings.

LikeLiked by 1 person
- 27
  
  hoakley on July 2, 2022 at 8:45 pm
  
  Thank you – yes, that’s a good way round when you’re stuck.
  Thankfully I very seldom have to dig my printer out and power it up now. Most of the time I just tell others that I don’t have one, and it’s amazing how they can be more flexible than they had been!
  Howard.
  
  LikeLike
28

Marc S on July 2, 2022 at 5:49 pm

Often when I want to save a web page I open the sharing menu in Safari, choose “Mail” and just save the mail without sending it. I then drag the draft mail to the appropriate folder – so I use Mail folders as (part of my) archive solution. Since I archive the majority of my (sent and received) mails anyway this just extends that.
I either need to be in Safari reader mode or I need to copy stuff from the website and manually paste into the mail.

However this has the disadvantage that Mail doesn’t save the image data within the mail, but only links. If images vanish from the website, Mail cannot show them anymore.
Thus sometimes I drag images from Safari to the Finder desktop, and then drag them from there into my mail. This will embed the real image data and not just a link.

If a website has lots of images, this is too much work. I then select the main body (trying to omit as much advertisement as possible) from the website, copy it (cmd-c), paste it into a new Textedit window, and save the file as .rtfd. This will change the formatting of the web page a bit, but usually the result is pretty good. I always drag the website address into the first line of the file (or mail) to have a link to the original web page right on top.
I prefer RTF over PDF because I can better edit the file (correcting typos, adding comments, deleting ads…), and the filesize is much smaller (especially after deleting unwanted stuff).

So the other half of my archive (outside Mail) is a bunch of folders with more RTF than PDF files.

LikeLiked by 1 person
- 29
  
  hoakley on July 2, 2022 at 8:54 pm
  
  Thank you.
  I agree in principle that RTF would be preferable to PDF, although it’s best for web page content which you’re going to edit, rather than save as a record.
  Howard.
  
  LikeLike
30

Rick Rump on July 2, 2022 at 9:48 pm

I appreciated this comprehensive-as-always article on saving webpages. Thank you for the work you put into compiling it.

LikeLiked by 1 person
- 31
  
  hoakley on July 3, 2022 at 12:29 pm
  
  Thank you.
  Howard.
  
  LikeLike
32

Jon White on July 2, 2022 at 9:58 pm

Nobody’s mentioned `wget`, which will download the HTML source and graphics to local storage (and, optionally, linked files).

I find it useful on occasion.

LikeLiked by 1 person
- 33
  
  Jon White on July 2, 2022 at 10:04 pm
  
  I appreciate, by the way, that this is off-topic for an article about saving web pages *from Safari*. I just mention it as a possibility for times when the four methods mentioned will not do.
  
  LikeLiked by 1 person
  - 34
    
    hoakley on July 3, 2022 at 12:30 pm
    
    Thank you.
    Howard.
    
    LikeLike
35

A passerby on July 3, 2022 at 8:42 am

Safari is a good browser but not suitable for archiving, even more on “modern” sites (layout problems, etc.)

The only “page save” use i’ve with Safari is to save pages as PDFs that I view in reading mode (pretty good for this browser). Even if it’s not PDF/A standard (useless for this use).

The webarchive format was still an interesting format a few years ago, but not anymore. And since we really don’t know how Apple wants to evolve with this format and its use in macOS and in WebKit…

Ideally, Apple should create a new simple, open and documented archiving format. Or at least it brings the missing elements in WebKit to support archiving extensions present in other browsers. Maybe one day…

—
[ Off Topic]
In the meantime, like others here, I use the SingleFile webextension with other browsers and the tools provided by Webrecorder (backup in .warc in particular, the website archiving format used by the Wayback Machine or National Library archivists) :

https://github.com/webrecorder/archiveweb.page

LikeLiked by 2 people
- 36
  
  hoakley on July 3, 2022 at 12:38 pm
  
  Thank you.
  Howard.
  
  LikeLike
37

Larry R on July 3, 2022 at 7:03 pm

Thank you for the thorough review of printing that has generated an interesting discussion. I have a few things to add.
1. Some web pages are already paginated even when displayed. In that case, using Export as PDF retains the page breaks.
2. When you use Export as PDF, the URL is kept in the metadata. It appears as ‘Where from’ if you do a Get Info on the pdf. Obviously, this is not included if you print to pdf.
3. The web site printfriendly.com offers a service that can create usable pdf’s from complex web pages. When you use it, it offers a preview that you can edit by deleting portions of the page you don’t want included in the pdf, such as advertisements. They have a browser extension (including for Safari) that makes it easy to use. It is a web service, so there are privacy implications, but they have a clearly written disclosure page on their site.

LikeLiked by 1 person
- 38
  
  hoakley on July 3, 2022 at 7:04 pm
  
  Thank you.
  Howard
  
  LikeLike