Process / 01—05
How it works
A measured path from archive evidence to a site ready to run.
Preview
Two paced archive index lookups, no page bodies. The preview shows capture counts, vintage range, redirect mix, and truncation.
Vintage
Restore as it was at or before a year, month, day or full timestamp. If you don't choose a vintage, the latest is used.
Captures per year · example
2016 · 171 captures
Restore
Wayback replay brings back pages first, capped at 500 pages and 2,000 assets. Common Crawl adds plain text for up to 100 pages Wayback missed. Requests are paced, so runs take minutes by design.
412 Wayback · 17 Common Crawl text · 9 missing
Viewer
Stored pages open with styles and images. Scripts stay off.
Unpushed
A finished restore is sent with a strategy. Unpushed creates the site.
Why it's paced
The archive limits how fast anyone can request pages. Going faster gets everyone blocked, including us. So runs take minutes by design.
What 'vintage' means
The site as it was at or before the date you pick. Enter 2019 and you get each page's latest capture up to the end of 2019. A page last captured in 2017 comes back as it was in 2017.
Where the pages come from
The Wayback Machine first. Common Crawl fills in text for pages Wayback missed, marked “text only”.