[submeta]

I created something like that for my spouse and myself. An app that creates an archive.ph url, extracts full text and generates summaries via an llm. I‘d open source it, but as I also extract paid articles via archive.ph, I think it wouldn’t be ethical to publish it.

[mstef]

the difference is, that archive.ph snapshots something in headless. omnom snapshots the exact same state that your browser is displaying you. so if there is js interactions that change the dom, those will be snapshotted, unlike with archive.ph.

also lets not forget that archive.ph wraps everything in their own frame and has their own way of mangling the result. not in a bad way, it's just not the original as it would have been rendered in your browser.

[viraptor]

Archive.ph is often used, because the "js interactions that change the dom" is typically a paywall. And you won't want it.

[mstef]

omnom is for snapshotting, not for circumventing paywalls. i'm merely comparing the snapshot feature of the two projects. circumventing paywalls is out of scope.

your bookmarks will never linkrot away.

[mstef]

btw it is perfectly fine to circumvent a paywall with archive.ph and then to snapshot it with omnon so your bookmark never linkrots away. also when i say "js manipulation" i also mean stuff like captchas, or dynamic documents that you change by interacting with it, or even private services like e.g. rocket chat hidden behind some barrier like http auth, or private vpn. archive.ph will never have access to what your browser might have access to.