Self-hosted Pocket alternative: save pages before they 404 with ArchiveBox
Pocket is gone. Mozilla retired it in 2025, and every list you carefully saved became an export file you had to find a new home for. But the deeper problem was there all along: a "saved" link in Pocket was still just a pointer. When the original page changed, went paywalled, or 404'd, your save rotted with it. Link rot quietly swallows the web. ArchiveBox is the opposite bet — it saves the actual page, in full, into an archive that lives on your disk and answers to you.
What ArchiveBox is
ArchiveBox is a self-hosted web archiving tool. You feed it URLs — one at a time, a bookmark export, an RSS feed — and for each one it captures a complete, offline copy of the page: the rendered HTML, a screenshot, a PDF, a single-file snapshot, the readable article text, even the raw WARC that professional web archives use. It's the same idea as the Internet Archive's Wayback Machine, running privately for you.
Read-it-later apps like Pocket kept a tidy list and, on their paid tier, a "permanent" copy on their servers, searchable for a subscription. ArchiveBox gives you the permanent copies and the full-text search with nothing metered and nothing to cancel. The archive is yours, it's plain files, and it doesn't disappear when a company changes strategy. This is the real archiving stack, not a bookmark manager wearing a trench coat.
The features that actually matter
- Every save is many copies. One URL produces a screenshot, a print-ready PDF, a self-contained single-file HTML, a cleaned "reader" version of the article text, a DOM dump, and a standards-compliant WARC. If one format ages badly, the others still open.
- Full-text search across everything you've saved. Search the contents of your archive, not just titles and tags, with a choice of backends (ripgrep, SQLite, or Sonic). This is the feature Pocket reserved for its paid Premium tier.
- Media, not just text. ArchiveBox drives yt-dlp under the hood to pull down video, audio, and whole playlists, so a saved talk or podcast is the actual file, not a dead embed.
- Import from where your links already are. Browser bookmark exports, RSS/Atom feeds, plain URL lists, and exports from Pocket, Instapaper, Pinboard, and Shaarli all parse cleanly — so migrating off Pocket is a single import.
- Save from the browser, manage from the web. The official extension adds a right-click "Save to ArchiveBox," so a page goes into the queue as you read it, and the built-in web UI — backed by a REST API and webhooks — lets you tag, search, and re-snapshot without touching the command line.
- Recurring, scheduled archiving. Point it at an RSS feed or a bookmarks URL and it archives new entries on a cron schedule, hands-off.
- Output that outlives the tool. The archive is just a folder of static files. You can browse it with any web server, or none at all, and it still works if ArchiveBox itself is ever gone. That's the whole point.
- Optional extras when you want them. Push a public copy to archive.org, clone the git repo behind a URL, mirror a whole site with wget, and follow a page's links one level deep to catch what it references.
Why self-host it instead of the hosted version
Your reading list is a profile. What you save says more about you than almost anything else, and a read-it-later company's real product was often that behavioural data and the "suggested" content sold against it. A self-hosted archive answers to no ad model. The pages you save are visible to you and no one else.
Cost that doesn't climb per seat. Pocket Premium was an annual subscription for the copies and the search; other read-it-later tools charge monthly per person. A self-hosted archive is a fixed slice of a server whether it holds a thousand links or a hundred thousand, and whether one person uses it or your whole team.
Control and permanence. You decide retention, which formats to capture, how long to keep media, and where the files live. Nothing gets deprecated out from under you, and nothing sends a surprise "we're winding down" email. When the vendor's strategy changes, your archive doesn't even notice.
Who it's for
- Researchers, journalists, and analysts who cite the web and need the source exactly as it was on the day they read it — not as it reads after a quiet edit.
- Developers and knowledge workers building a personal knowledge base, who want saved docs, threads, and posts searchable and permanent instead of a graveyard of dead tabs.
- Anyone leaving Pocket with years of saved links, who would rather import the whole list once and own the copies than go shopping for the next app that will also shut down.
The honest setup notes
Rolling this yourself is real work, and we'd rather just say so.
- Archives grow fast, especially with media. Text snapshots are tiny; a saved playlist of videos is not. Disk is the resource you actually plan around, and you need a real backup of the archive folder and its SQLite index — the copies are only "permanent" if they're backed up.
- Archiving is memory-hungry. ArchiveBox runs a headless Chromium to render screenshots, PDFs, and the DOM, which spikes RAM while a batch is processing. A too-small box will stall on big saves.
- Updates and extractor upkeep. New releases, and the underlying tools like yt-dlp, need periodic updates to keep capturing sites that constantly change their markup.
- Some pages fight back. Sites behind a login, a hard paywall, or aggressive bot protection archive incompletely. You can hand ArchiveBox your session cookies to get past a login, but it's fiddly, and a few pages simply won't capture cleanly — no archiver dodges that entirely.
None of that is exotic, but it's a standing chore. Or let us run it — we handle the disk headroom, the backups, and the updates, and you just save links.
Deploy it in one click
Self-hosting ArchiveBox the hard way means a Docker host, a volume big enough for the archive, Chromium dependencies, a reverse proxy, TLS, and a backup job you actually test. It's very doable, and it's an afternoon you don't get back.
On Caliber Node it's one click. Pick ArchiveBox from the catalog and we provision it on managed infrastructure with storage, HTTPS, and updates handled — and you still get real access underneath: full logs, SFTP to your files, and an archive you can export any time. You can deploy ArchiveBox in one click and point it at that Pocket export before it gets buried in your downloads folder.
First run
- Open your ArchiveBox URL and sign in, then set the admin account so the web UI is yours.
- Grab your bookmark export — the Pocket file you saved before it closed, or a fresh export from your browser.
- In the UI or CLI, add that file as an import; ArchiveBox parses the list and queues every URL.
- Watch the first snapshots complete — each row fills in with a screenshot, PDF, article text, and the rest as the extractors run.
- Search for something you saved years ago and open the offline copy. That page is now yours, 404 or not.
The bigger picture
Owning the tools you rely on isn't nostalgia — it's the difference between a library and a rented shelf. The web you read is being quietly edited, paywalled, and deleted, and the companies that promised to "save it for later" keep closing the tab for you. An archive you host is a small act of permanence: the pages you thought were worth keeping are still there, in full, searchable, on your terms.
Deploy ArchiveBox in a few minutes →
One-click self-hosted apps, managed for you — real logs, SFTP, and your data always exportable. Never a black box.