Why websites disappear from the Wayback Machine (and how to actually get yours back)

If you are reading this, you probably lost a website. Maybe the hosting lapsed, the developer vanished, or the domain expired and got scooped up. You checked the Wayback Machine and there it was — your old site, still visible. Relief. Then you tried to “save” it, and reality hit.

The Wayback Machine is not a backup

The Internet Archive crawls the web on its own schedule. It does not mirror your site. It takes snapshots — sometimes months apart, sometimes never for the pages that mattered most. What you see on web.archive.org is the intersection of “when the crawler happened to visit” and “what the crawler was allowed to capture.” That intersection is almost never your full site.

• Crawl gaps. A blog that ran from 2018 to 2023 may have 40% of its posts archived, the rest never crawled.
• Asset drops. Images, fonts, CSS and JS are frequently missing because they lived on CDNs the crawler rate-limited.
• Parametric URLs ignored. Pages like ?p=42, ?id=7, /page/3/ are often skipped because the crawler treats them as duplicates.
• Robots.txt after death. When a domain changes hands, the new owner’s robots.txt can retroactively hide old snapshots.

What “restoration” usually means on Fiverr

Open Fiverr, search “wayback restore,” and read the fine print on the $5–$30 gigs. A pattern emerges quickly:

Most budget sellers deliver a static HTML dump. WordPress is quoted as an extra. Parametric URLs are “not included.” Broken images are your problem.

That is not restoration. That is a screenshot of your site, wrapped in HTML, handed back to you to fix. You then pay someone else — or spend a weekend — turning that dump into something editable and live. The advertised price is a floor, not a total.

What a real restoration looks like

A proper recovery has to solve three problems at once: capture, cleanup, and relaunch. Skip any one and you get a half-working site.

1. Capture — every URL, including the weird ones

This is where my own tooling earns its keep. I do not save pages one by one. I run a full-site capture that walks the archive and recovers parametric URLs (?p=1, ?id=42) as distinct pages — the exact thing most tools quietly skip. For a catalog or listings site, this is the difference between a usable restore and a useless one.

2. Cleanup — remove the archive, not the content

Every archived page ships with junk: the Wayback toolbar, analytics beacons, dead tracking pixels, links pointing back into the archive (/web/20210101/…), the previous owner’s contact details, and footer dates from 2014. AI-assisted cleanup strips all of this in one pass and refreshes the footer to the current year across every page — including the ones with thousands of URLs where a manual pass is impossible.

3. Relaunch — a working WordPress site, on your domain

The cleaned pages go into a real WordPress install as full HTML blocks. Not Gutenberg rows you have to rebuild, not a static folder you have to host separately. Domain, DNS, SSL, Cloudflare, sitemap, and a 404→301 redirect rule that stops dead links from hurting SEO. You get admin access on day one.

How to check your own site before paying anyone

Before you commission a restoration, do a 5-minute self-audit so you know what you are buying:

• On web.archive.org/web/*/yourdomain.com, note the snapshot density. A site with 12 captures in 2022 is recoverable. A site with 2 captures total is partially recoverable at best.
• Click into 3–4 inner pages, not just the homepage. If they 404 in the archive, they will 404 in the restore.
• Check whether the archive captured ?p= or ?id= URLs. If your site used them, list them — they drive the price and the scope.
• Right-click an image and open it. If it is broken in the archive, it will be broken in the restore unless you supply a replacement.

The honest version

Not every site can be fully recovered. Some are gone for good — the snapshots simply do not exist. But far more sites are recoverable than the $15 Fiverr gig implies, and far fewer are “done” than the seller claims when they hand you a folder of HTML. The gap between looks recovered and actually works on your domain is where most people lose money.

If you have a domain and a Wayback link, send it over. I will tell you honestly what is recoverable, what is not, and what it costs — before any work starts.