Recovering a 20-year blog from the Wayback Machine
When the server behind blog.wo.ai died with no backups, I rebuilt all 316 posts and their images from Internet Archive snapshots.
- archive.org
- recovery
- static site
- Astro
In 2025 the server behind blog.wo.ai died, and the database and uploads went with it. The writing went back to 2005.
The Internet Archive had crawled the blog for years. The recovery went like this:
- List every captured URL on the domain through the Wayback CDX API.
- Download 889 raw snapshot pages, picking the best capture of each post.
- Pull out the title, date, categories and body of each post, and convert them to Markdown with front matter.
- Download every image the snapshots referenced. A group of 2010–2011 photos had been hotlinked from QQ Zone through an anti-hotlink proxy, and they were still on QQ’s servers when requested with a browser User-Agent.
- Rebuild everything as a static Astro site that keeps the original
/archives/<id>permalinks.
The result: 316 posts and 177 images, now also published here on the blog.
Comments