The service shut down.
A free bookmark feature is a cost centre attached to somebody else's strategy. Sooner or later the strategy changes, and you get ninety days' notice and a zip file.
The short version
- This category closes more than most. Bookmark and read-later tools are cheap to launch, rarely profitable, and easy to shut down.
- An export button is not portability. A file you cannot import anywhere useful, with the metadata stripped, is a souvenir.
- Portability is four things: an open format, a documented protocol, source you can read, and a server somebody other than the vendor can run.
- MarkSync is built so it can outlive us. GPL-3.0 code, a documented API, and self-hosting as a first-class option rather than an enterprise tier.
A short history of disappearing bookmarks
If you have been online for twenty years, you have lost bookmarks at least once. Social bookmarking services were a genuine movement in the 2000s, and the best known of them changed hands repeatedly before fading. Google ran a bookmarks product for years and retired it in 2021. Read-later services, which are bookmarks with better manners, have come and gone on a similar cycle. Browser vendors have quietly removed sync features and changed account requirements along the way.
The pattern is not incompetence. It is structural. Storing links for people costs money, earns almost nothing directly, and competes for engineering attention with whatever the company actually sells. When the strategy shifts, the bookmark product is an obvious thing to cut, and the people who cared most about it are the smallest group affected.
Why the export button does not save you
Shutdown announcements almost always come with an export. It is better than nothing, and it is much less than it sounds.
- It is often lossy. Tags, notes, highlights, archived copies, folder structure and dates are the parts that took you years to build, and they are the parts most likely to be flattened into a list of URLs.
- It arrives in a format nobody imports. A proprietary JSON dump is not something your next browser has ever heard of. You will be writing a conversion script, at the precise moment you have other things to do.
- It is a snapshot with a deadline. You have until the lights go off, and if you miss the window, there is no second chance.
- It does not restore the workflow. You can get the data out; you cannot get back the thing that put it there, which is what you actually used every day.
Lock-in rarely looks like a locked door. It looks like a door that opens onto a field.
What portability actually requires
"Can I export?" is the wrong question. The right one is: if this company stopped existing tomorrow, could someone else keep this running? That needs four things, and most services have one or two.
An open data format
Your bookmarks should be expressible in something other tools already read. For bookmarks there is a de facto standard — the HTML bookmark file that every browser imports — and anything you cannot get into that shape is a format with a hostage in it.
A documented protocol
If the client and the server talk over an API that is written down, anyone can write a new client, or a new server. If the protocol is undocumented, the client is the only way in, and the client's days are numbered with the company's.
Source you can read
An open licence turns a product into something closer to infrastructure. If the company walks away, the code does not: a fork can continue. MarkSync itself exists because of exactly that — it is a maintained fork of xBrowserSync, which is only possible because that project was open in the first place.
A server you can run
This is the one that is usually missing. Self-hosting is frequently a marketing tier rather than a real option. If running the service yourself requires a licence negotiation, you do not have an exit; you have a brochure.
How MarkSync is set up for this
MarkSync is a fork of xBrowserSync, kept going by volunteers, and it is deliberately arranged so that no single party — us included — is load-bearing.
- The code is GPL-3.0. All of it, on GitHub. If this project stalls, somebody else can pick it up, exactly as we picked it up.
- The API is documented and stable. The extension, the web app and every public service speak it. It is the same protocol xBrowserSync used, which is why an existing sync keeps working untouched.
- You can host it yourself, and that is the recommended path for anyone who takes this seriously. Not a paid tier, not a pilot programme — just the software, running on your machine.
- You are not tied to one service. Several public services run the same API, so if one operator goes quiet you can point the extension at another, or at your own, without changing tools. The current list is on the home page, with live status read straight from each one.
- Your data is portable by construction. It decrypts on your device, and what you get back is your bookmark tree, not an archive you need a parser for.
And the unglamorous part: because your bookmarks are encrypted with a password only you hold, nobody can recover them for you — not us, not the service operator. Portability and recoverability pull in opposite directions. Keep a backup; the extension will make you one.
What to do this week, whatever you use
None of this is a reason to panic about whichever service you are on today. It is a reason to spend ten minutes on three things.
- Export now, not when you need it. Find the export, run it, and check that the file actually contains your folder structure.
- Check whether the server is something you could run. That single question tells you how much of your setup is a promise and how much is a property.
- Prefer a stack with more than one operator. One company with one server is a single point of failure, however well run it is.
If you want the reasoning behind picking a tool that cannot read your data in the first place, that is what a bookmark collection says about you.