How to archive a web page for the long termWayback Machine vs. archive.today vs. ArchiveWeb.page vs. Page Scanner

To keep a web page for years, you have three kinds of copy to choose from. A public archive, such as the Wayback Machine or archive.today, holds a copy on its own servers that anyone can look up by address. A web archive file, such as the WACZ that Webrecorder's ArchiveWeb.page writes, is the page itself, scripts and all, in a file you keep and replay later. A PDF/A is a document: fixed, printable, readable in any PDF reader, and made to a standard built for long-term storage. They answer different questions, and for anything that matters you may want two of them.

Four tools at a glance

✓ yes, ✗ no, "signed in" where the feature needs an account. An empty cell is one the product's own pages do not answer.

What PDF/A is

PDF/A is ISO 19005, the PDF standard for archiving. Part 2 builds on PDF 1.7, and its basic conformance level, "b", is about appearance: the file has to look the same when it is opened years from now, while the stricter level "a" also requires tags that describe its structure. So the file carries what it needs for that: its fonts embedded, its color space named, and metadata an archive can check, which a validator such as veraPDF tests it against. A PDF/A of a web page is a document, not a page: links can stay clickable, but nothing on it runs. That is the trade against a web archive.

The Wayback Machine (Save Page Now)

The Internet Archive's form puts it in one line: "Capture a web page as it appears now for use as a trusted citation in the future." Paste an address, press the button, and the page is saved to the Wayback Machine, where "anyone can access our collections". Save outlinks is on the form for anyone; a screenshot, a copy in your own web archive and a WACZ of the results by email are listed under "Sign in to use extra features".

Its strength is that someone else holds the copy: a Wayback link in a footnote opens for a reader who has never met you. The costs follow from the same fact. The Archive says it does "not archive pages that require a password to access", so a page you are signed in to is out of reach. Scripts that need the original site do not come along: "when a dynamic page contains forms, JavaScript, or other elements that require interaction with the originating host, the archive will not contain the original site's functionality." And a site owner can have pages excluded by request.

Verdict: the first thing to do with any public page you might cite. Free, public and held by a third party.

archive.today

archive.today calls itself "a time capsule for web pages". It "saves a text and a graphical copy of the page", the graphical one being a "screenshot of 1024x768 pixels", and gives each capture a short link. Its FAQ is plain about what it keeps: text, images, frames and content generated by JavaScript, but not video, sound or PDFs, and a stored page "must be smaller than 50Mb". There is no account; the FAQ answers "Can I have an account" with "No."

Saved pages "will have no active elements and no scripts", which the site presents as safety. Its description of what a capture means is modest: it "merely certifies that at the given point of time there was a page on the web." The FAQ also says that when you archive a page, your IP address is passed to that website in an X-Forwarded-For header.

Verdict: a quick public copy with a short link, and a picture of the page beside the text.

ArchiveWeb.page

Webrecorder's ArchiveWeb.page is a Chrome extension and a desktop app for "interactive, high-fidelity web archiving that runs directly in the browser". You start a capture and browse; it records what the browser loads as you go, and its Autopilot "will scroll or interact with certain complex sites", such as infinite-scroll feeds. Archives are kept in the browser's own storage, and you download them as WACZ or WARC.

WACZ is Webrecorder's format for "web archive collections to be packaged and shared on the web as a discrete file": the recorded WARC data, indexes and a list of pages in one zip. Opened in ReplayWeb.page, it is the page again. Webrecorder's words are "browse fully interactive archives", and the extension saves "complex interactions as you browse": that is what a PDF cannot keep.

What it costs: the file opens in a replay tool rather than a PDF reader, and it is a copy you hold rather than a public one. During a capture Chrome shows "started debugging this browser", as it does for Page Scanner (here is why).

Verdict: the most faithful copy of a page that does things. Pick it when behavior matters as much as appearance.

Page Scanner: PDF/A and an integrity record

Page Scanner captures the page in your own Chrome, signed in or not, and exports it from the editor. In the Output Format panel, under Archiving, PDF/A-2b, for archives turns the ordinary PDF into a PDF/A-2b: the fonts it names are embedded, sRGB is named as the output intent, the XMP metadata declares PDF/A-2b, and the file gets an identifier. The page's own text stays selectable, its artwork stays vector, and its links stay links. Every test file we made this way, Korean text, a redacted page, a page at three widths and two tabs in one file among them, passes veraPDF as PDF/A-2b.

The limits: it is the basic level, -2b, which promises the file looks the same, not that its text extracts exactly (-2u) or that it is tagged for accessibility (-2a). It can be up to half a megabyte larger, for the fonts. A text mark in Serif, Mono or italic becomes a picture of its words, still selectable. PNG and JPEG have no archival form, and an export through the command line or an agent is always the ordinary PDF.

The Integrity panel adds an optional <file>.integrity.json beside the export. It records the file's name, size and SHA-256; the page's address and title; a SHA-256 of the page's HTML as the scan found it (not the HTML itself); when it was captured and exported, in UTC and in local time; the capture mode, crop and format; the extension's version and the browser. The record, signed also signs it with a key made in this browser, and the panel shows the key's thumbprint.

To check a file later:

npx @page-scanner/cli verify terms.pdf

verify recomputes the file's SHA-256 and length, checks the signature and that the key matches its thumbprint, and says that the time in the record is the capturing computer's clock. What that shows is narrow, and the record says so in itself: the file is byte for byte the one exported, and, signed, whoever holds that key made the record. It does not prove when the capture happened, because the time is your computer's clock. It does not prove what the page showed anyone else, or what the server sent. The key cannot be backed up or moved to another browser, though records it signed still verify with the public key they carry.

Captures, marks and files stay on your computer; the extension sends anonymous counts of which features are used, and you can turn that off in the settings.

Verdict: the one to use when the archive has to be a document: a page behind a login, a file for a records system, or a copy you will print or attach.

Which should you use?

Use the Wayback Machine if the page is public and you want a copy someone else holds, one a reader can open from a link without trusting you.

Use archive.today if you want a public copy with a short link and a screenshot beside it.

Use ArchiveWeb.page if the page's behavior is part of what you are keeping, and a reader should be able to click through it later.

Use Page Scanner if you need a file in a document standard: PDF/A-2b, readable in any PDF reader, made from the page as you see it, with a digest to check it against.

Use two when it matters. A Wayback capture and a PDF/A of the same page are different kinds of record: one is public and held by a third party, the other is yours and opens without a network. Web clipper or page capture? makes the same case for notes.

Frequently asked questions

Is a PDF/A proof of what a page said? No. It is a file format for keeping documents, and it promises appearance, not origin. Page Scanner's integrity record shows that a file has not changed since it was exported; it does not show when, or what anyone else saw.

Can the Wayback Machine save a page I am signed in to? The Internet Archive says it does not archive pages that require a password. Page Scanner captures the tab in your own browser as you see it, signed in, and the file stays with you.

Why not keep the HTML? You can, and a WACZ is the most complete way to do it. A PDF/A answers a different need: a fixed document that any PDF reader opens the same way. For the plain "save as PDF" question, see how to save a web page as a PDF.

Where this comes from

• Internet Archive: the Save Page Now form, the help center's Using the Wayback Machine, Wayback Machine General Information and Save Pages in the Wayback Machine.
• archive.today: the home page and the FAQ.
• Webrecorder: ArchiveWeb.page, its capture guide and download guide, the repository, the WACZ specification, ReplayWeb.page and webrecorder.net.
• veraPDF, the open-source PDF/A validator.
• Page Scanner: Archiving, An integrity record and page-scanner verify.