You open your old site in the Wayback Machine and the text is there, but the logo is a broken icon, the menu is a bulleted list and the layout looks like a document from 1997. Nothing is wrong with your browser. The archive kept some of the pieces and not others.
Why archived pages come back broken
The Internet Archive’s FAQ is plain about it: “Broken images occur when the images are not available on our servers” (Internet Archive Help Center). In practice that means the crawler never captured those files. The same page says that “simple html is the easiest to archive”, and that JavaScript is hard to archive. It also lists reasons a page may be missing altogether: the crawlers did not know the site existed at the time, it was password protected, or it was blocked by robots.txt.
So an archived page is not one snapshot. It is an HTML file plus whichever images, stylesheets and scripts happened to be captured, each possibly on a different date.
That is the key point, and it explains most of the damage. The crawler saves the page’s HTML on one visit. The images, the stylesheet and the fonts are separate files with their own web addresses, so they are saved on their own visits, if at all. When a stylesheet was never captured, the browser has no styling to apply, so it falls back to its plain defaults. That is the “1997” look: black text on white, underlined links, bullet lists. The content is intact. Only the instructions for how to display it are gone.
Images fail the same way. The page still points at the old image address, but if the archive has nothing stored there, you get the broken icon. A background image set in CSS disappears for the same reason the whole stylesheet might: it was a separate file.
What you can check yourself
- Open the page at several dates. A page from one month may have its images; the same page from another month may not.
- Look at the missing file’s address. If the stylesheet or image has a capture on any date, it can be reused even if the page’s own date lacks it.
- Turn JavaScript-heavy parts into a note. Sliders, forms and search boxes usually do not survive. Plan to rebuild them, not recover them.
- List what is truly gone. Anything never captured cannot be restored from the archive. Write it down early so it does not surprise you later.
To find a missing file’s address, open your browser’s view-source or developer tools on the archived page and read the src and href values. Those are the files the page is asking for. You can then paste each address into the Wayback Machine on its own to see whether it was ever captured, and on which dates. This is slow by hand, but it tells you the truth about what survives and what does not.
The free Wayback Press scanner shows the old design, the URL structure and a few sample pages, so you can see this before you spend any time.
What the archive can and cannot give back
It helps to sort the pieces before you start, because the archive treats them differently.
- Usually recoverable: page text, headings, links, and any image or stylesheet that was captured on some date.
- Sometimes recoverable: the exact look, when the stylesheet survives on a nearby date even if not on the page’s own date.
- Rarely recoverable: anything driven by JavaScript, such as sliders, live search, comment widgets and contact forms. These run code that the archive does not execute, so they come back dead or empty.
- Never recoverable from the archive alone: files that were behind a login, blocked by robots.txt, or simply never crawled. If the Internet Archive never saw it, no tool can pull it out.
Email, databases and anything served only to logged-in users are not in the archive at all. The Wayback Machine keeps the public, front-of-house version of a site, not its back end.
Three ways to deal with it
| Option | Good when | Limit |
|---|---|---|
| Accept the gaps, rebuild the look in a clean theme | The text matters more than the old design | The site will not look identical |
| Pick the best snapshot per file by hand | A handful of pages | Slow once pages run into the dozens |
| Let an LLM-checked pipeline choose snapshots and fix layout breaks | Many pages, mixed eras | Still limited to what was archived; a human should review the result |
The third row is what HTML to WordPress conversion here does: my agent picks the most complete snapshot per page, an LLM checks the design, and a human reviews before delivery. If you would rather run it yourself, the Wayback Machine downloader is free and has no AI.
Common questions
Can you force the archive to re-fetch a missing image? No. You can ask the Wayback Machine to capture a page that is still live today, but you cannot make it capture a file from the past. What was not saved then is not there now.
Why does one page look fine and the next looks broken? They were captured on different visits, and they may depend on different files. A page that happened to be crawled when the stylesheet was also available will look right; its neighbour may not.
Can the look be rebuilt if the stylesheet is gone? The exact original cannot be recovered without the original file. The look can be rebuilt by hand or in a clean theme, using the surviving pages as a reference. That is a rebuild, not a recovery, and it will be close rather than identical.
Limits
- The archive is the ceiling: files that were never captured cannot be converted.
- You must have the right to use the content you restore. See the terms.
- Not affiliated with the Internet Archive or its Wayback Machine.
- No ranking promises: a restored site is a starting point, not a result.
If the whole site is gone and not just its styling, start with recovering a deleted website.
The archive keeps what it was shown. It does not keep what you meant.