Posted by ingve 6 hours ago
That ArchiveWeb extension looks excellent for custom captures. Picking the pages and elements you want during capture is a great feature.
For quicker one-offs, SingleFile gives you a single HTML with all the media of a page: https://www.getsinglefile.com
If you're making archiving a more involved part of your workflow, I highly recommend ArchiveBox. It can save pretty much every type of file from everywhere, in almost every format you can imagine, while building a searchable index of the archives. Been using it successfully for many years: https://archivebox.io
Link rot really bothers me and I want people to get into archiving web pages so that eventually there will be a decentralized mirror of the unslopped web sitting on people's hard-drives, ready to share whenever someone needs it. So, I made a worse version of ArchiveBox ( https://memoryhole.app ). My philosophy is that, if people see how it is easy to start archiving and how useful it is, then they'll keep it up with SingleFile, or ArchiveBox or anything else.
I did small amount of troubleshooting a few times with no resolution, but now I'm thinking of switching my thought process. Maybe something like a "web clipper" is much better, as complex SPA pages often fail to archive well anyway. And I should focus on content itself, not the page. That SingleFile extension fits well here as well.
Did you run into similar dilemmas at some point?
At this point I'm using Linkding (self-hosted bookmarks, which has an archive function) paired with Singlefile (which integrates with Linkding) which locally captures and uploads. For larger, or multi-page, captures I still fire off an Archivebox command. I run into blocks occasionally, but it also just feels like the open web is becoming less and less open in general.
My bookmarks file is a graveyard as I have 20 yrs’ old ones…
And as a publisher I don’t link fresh sites (don’t even mention Netlify, Lovable, Vercel…) … Vannevar Bush would be surprised how his Xanadu evolved.
Please, please buy books and keep them in good condition, and please archive websites with important information (assuming it's legal in your jurisdiction). There are still plenty of excellent sites and blogs around from the 2000s and even earlier, but they won't be there forever. Ironically, coding agents have made archival much easier -- I've always found problems with using pre-written tools, but Claude or Qwen make it trivial to download, test, and archive sites on your local.
Hacker News has sure dropped off.
If you mean 'an LLM' as in the owner/maintainer of the LLM/weights, then examples of this already exist. Whether it's DeepSeek/Gwen downplaying Uyghur forced labor camps [0] or Grok hallucinating a 'white genocide' in South Africa [1], it is clear that the owners of LLMs have political agendas that are expressed in the weights of popular LLMs.
On the other hand, if you mean 'an LLM' as in an autonomous agent that could decide to rewrite history, that also seems very possible. Either by using existing techniques like, "poisoning ChatGPT, Gemini, and Google AI Overview answers by seeding the Web with malicious links and data and then optimizing the content." [2] Or by breaking out of its sandbox and updating its own internal weights that allow for an outcome that distorts historical facts. "In July 2026... OpenAI models circumvented controls designed to isolate them from the internet and compromised parts of OpenAI’s internal research infrastructure and Hugging Face’s systems ." [3]
[0] - https://www.cbsnews.com/news/china-ai-qwen-artificial-intell...
[1] - https://www.cnbc.com/2025/05/17/groks-white-genocide-respons...
[2] - https://www.darkreading.com/threat-intelligence/attackers-ma...
[3] - https://openai.com/index/hugging-face-incident-and-the-road-...
It had lots of photos that I hand-scanned from printed film, back in the day when it took a good bit of time to scan photos manually.
The site was a running joke, where I would open it to show people that it still existed... until it didn't, and I couldn't find my backup of all of the photos. The archive sites have the 'bones' of the website, but not the photos. It was easily avoidable, which is why it is painful.
I guess the web plugins could ideally get around it, but what about the CLI and app solutions?
I know yt-dlp has gone to great lengths to work this problem.
I thought it might be rose-tinted glasses, thinking that the earlier internet was better, but spending time retracing my teenage steps made me even more certain that it was indeed better. So many derpy little websites made by nerds, so many fun games and websites that weren't preying on vulnerable people, and everything spread out across many places - not centralised in the hands of large corporations. I've been using LLMs to have fun recreating concepts from that time, like my own forum software, as some sort of anti-progress. It's been fun.
(Now if hard drives had not become so damn expensive—it's getting harder to be a digital hoarder like myself.)
You don't have to worry about having your copy easily accessible, and as a bonus it's also available for everyone else in the world.
i managed to reconstruct (and then improve) it with an LLM:
History and this very thread should show you how this is a naive statement.