{
"$type": "site.standard.document",
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreih4xssxz3d2a542svzdpsj7hp7ylsxhwqsrs6bp2nydvvrrm3oc5e"
},
"mimeType": "image/png",
"size": 231834
},
"path": "/notes/claude-code-to-pull-old-articles-into-astro/",
"publishedAt": "2025-12-18T00:00:00.000Z",
"site": "at://did:plc:aes3lokiqtv63fk62nwnjeuf/site.standard.publication/3mnin5cnq2q2a",
"textContent": "I read Simon Willison's blog religeously, and occasionally he references a post he wrote way back in the dark ages of 2005. Every time he does I think what a shame it is that I lost a lot of the content I've posted over the years by intermittently switching the platform and domain I publish on. I've various published on: - http://danny.is (since about 2013) - http://dasmith.co.uk - http://blog.dasmith.co.uk - http://notes.dasmith.co.uk - http://thescri.be In July I stumbled on a GitHub repo which had some of my old posts from dasmith.co.uk, and ported them over to this site. Today I spent a little bit of time looking on the Wayback Machine to see if I could find and move over some of my much older writing - particularly the stuff I published on my first blog (thescri.be) circa 2005-07. I asked Claude Code to: It came up with a plan which involced using the chrome-devtools MCP to spin up a browser and navigate the internet archive to find as many posts as it could from various snapshots. It did a really good job of finding posts which only appeared in one snapshot, which it turned out was quite a lot of them (all in different snapshots). It then extracted their contents and downloaded any archived images and created some markdown files with YAML frontmatter for them. I then asked it to: So in two prompts (three if you include after it did the planning) I ended up with a load of my super old writing as notes in this site! I'll do something similar for the other old domains I have soon!",
"title": "Using Claude Code to pull old articles into this Astro site"
}