X bookmark folders: what Premium actually gives you

August 12, 2026

Manual, flat, and Premium-only. Useful for curation, not for a backlog.

Folders come with X Premium. They are named containers you put saves into by hand, one at a time. There is no automatic sorting, no rules, no tags, and no nesting. If you are on a free account you do not have them, and no setting will reveal them.

Setting them up

Open Bookmarks from the side navigation.
Use the add-folder control on the bookmarks screen and give the folder a name.
To file a save, open the overflow menu on the bookmarked post and choose the folder.

You can also pick a folder at the moment you bookmark something, by holding the bookmark control rather than tapping it. That is the version worth learning, because filing at save time is the only way folders stay current without a weekly tidying session.

What they do not do

This is the part that decides whether folders solve your problem:

Nothing is automatic. No sorting by author, topic, keyword, or anything else. Every post lands in a folder because you put it there.
A post lives in one place. Folders are containers rather than labels, so there is no equivalent of tagging one save as both "AI" and "hiring".
No nesting. One flat level of folders.
They do not raise any limit. The retrieval ceiling that stops a large collection paginating applies the same either way.
They are not an export. Folder structure is yours inside X and comes out only through third-party tools, if at all.

Filing at save time versus filing later

There are two moments you can file something, and they are not equally good.

At save time you still know why you kept it. The thread is on screen, the context is in your head, and choosing a folder costs about a second. This is when filing is nearly free and nearly always correct.

Later you are looking at a link with no memory of the conversation it came from. You have to reconstruct why past-you cared, which takes ten times longer and produces worse answers. This is why bulk-filing sessions feel so grim: you are not sorting, you are doing archaeology on your own attention.

The practical upshot is that folders are worth adopting the day you decide to use them and are close to worthless applied backwards. Start filing new saves now and leave the old ones alone.

Folders are not tags, and the difference matters

A folder is a place a thing lives. A tag is a label a thing carries, and one thing can carry several. X gives you the first and not the second, which sounds like a detail and decides how you have to think.

With tags, a post about hiring engineers at an AI startup gets three labels and turns up in all three searches. With folders you have to pick one, and the picking is where the system breaks down: you hesitate, the hesitation costs a second you did not want to spend, and the post goes unfiled. Do that for a month and the unfiled pile is bigger than all the folders combined.

The workaround people reach for is folders named after combinations, which is how you end up with "AI hiring" and "AI tools" and "Hiring" and no idea which one anything is in. Resist it. Fewer, broader folders lose you some precision and cost you nothing in hesitation, and hesitation is what actually kills the habit.

A folder set that actually gets used

If you want something concrete to copy, this holds up across a lot of people:

This week. Things you have genuinely committed to reading in the next few days. If it is still there next week, it was not that.
Reference. Things you expect to look up again rather than read: docs, a good explanation of something, a link you will need.
Reply. Things needing a response from you. These are time-sensitive in a way nothing else in the list is, and mixing them in with reading material is how you miss them.
Someday. One honest dumping ground, which you should expect to delete wholesale once a year without reading. Having it named keeps the other three clean.

Four folders, all named after what you will do rather than what the thing is about. Notice that none of them is a topic: topics tell you what a post is, and the question you will actually have in a month is why you kept it.

What happens if you stop paying

Worth knowing before you invest hours in filing: folders are a Premium feature, so losing Premium means losing access to the organising layer. Your bookmarks themselves are not the subscription; they stay saved. The structure you built on top of them is.

That is a reasonable argument for keeping folders broad and few. A handful of coarse buckets is easy to reconstruct anywhere. A carefully maintained twenty-folder taxonomy exists only inside a subscription, cannot be exported by X, and evaporates the month you decide to cancel.

Are they worth it for a backlog?

Honestly, no, and it is worth being clear about the arithmetic. Filing a save takes a few seconds: open the menu, pick the folder, confirm. Call it five seconds with the thinking. At two thousand saves that is nearly three hours of uninterrupted filing, and it produces a tidy archive of things you still have not read.

Folders earn their keep going forward, on saves as they arrive, where the cost is a second each and the benefit compounds. They are a curation tool, not a recovery tool.

A folder scheme that survives

Two rules, both learned the hard way by anyone who has tried this:

Few and broad. Four or five folders get used. Fifteen specific ones turn every save into a filing decision, and a filing decision you cannot make instantly is a save that goes unfiled. The unfiled pile then becomes the real bookmark list and the folders become decoration.

Name them by what you will do, not by what they are about. "Read this week", "Reference", "Reply to" beat "AI", "Design", "Startups". Topic folders tell you what a thing is; action folders tell you why you kept it, which is the thing you will have forgotten in a month and the thing that determines whether you ever open the folder again.

If you are deciding whether to pay for Premium for this

Folders alone are a thin reason. The search box that comes with the same subscription is the more useful half, because it works without you having pre-sorted anything and it helps most in exactly the situation where folders help least, which is a large existing collection.

If the collection you are trying to tame is already thousands deep, neither one fixes it. What fixes it is the number going down.