Hide rage-bait on X without leaving X

What "rage-bait" actually is, and why X serves it

Rage-bait is content engineered to make you angry enough to engage — quote-tweet, reply, share with a snarky line. Engagement-optimised algorithms love it because it lifts every metric they look at. The X For You tab is engagement-optimised. The two facts compose.

The mistake most people make is treating rage-bait as a political problem. It is not; it's an emotional problem. Plenty of rage-bait has no politics in it — fake productivity claims, manufactured drama between strangers, reply-baiting "hot takes" on topics nobody actually disagrees about. Filtering only by topic misses half of it. Filtering by emotional valence catches it.

The cheapest fix: stop using the For You tab

The single highest-leverage thing you can do on X in 2026 is stop opening the For You tab. The Following tab — chronological posts from accounts you actually follow — is on the same screen, one tap away, and is not engagement-optimised in the same way. You'll see fewer posts; you'll see fewer of them designed to manipulate you.

If you only do one thing from this post: set Following as your default. The toggle is persistent on the web client. On mobile, you have to tap it each session, which is mildly annoying but works.

This is the equivalent of switching from Reddit's home feed to your subscriptions feed — same logic, same result.

X's built-in filters: useful, but limited

X gives you three native tools:

• Mute keywords: Settings → Privacy and safety → Mute and block → Muted words. Add the obvious offenders (specific public figures, recurring rage-bait phrases). Set durations — "30 days" is usually right; if you forget about a muted phrase, that's a sign it didn't matter.
• Mute accounts: from any tweet, the three-dot menu. Use this generously. Muting is not blocking; the muted person never knows. It's not a moral judgment, it's an inbox cleanup.
• Lists: build a private list of the 30–60 accounts you actually want to read, pin it as a tab. This is the manual version of the For You algorithm, and it's almost always better than the auto-curated one.

These three together cover maybe half the noise. They miss anything that doesn't keyword-match or come from a person you've already chosen to mute.

Algorithmic-level filters: blocking topics you haven't seen yet

X's word filter is reactive. You see something annoying, you mute the phrase, it stops appearing. Anything you haven't pre-muted gets through.

A content classifier — like PureFeed for Pro users — is proactive. It reads the actual text of each post and scores it on dimensions like sensationalism and negativity. You set a threshold ("hide if sensationalism > 75") and it catches new rage-bait the moment it appears, regardless of which words it uses or who posted it.

The difference matters most for the long tail. Anyone can mute the five accounts that produce the most rage-bait. Catching the next ten thousand accounts that produce occasional rage-bait without you ever having to learn their names is the actual problem.

A practical setup for "I want X, just less inflamed"

The combination that works for most people:

• Default to the Following tab; touch For You only deliberately
• Mute aggressively — both keywords and accounts. Re-evaluate quarterly; mutes accumulate
• Build one Following list of 30–60 accounts you genuinely want to read; pin it as a tab. Use it as your primary feed
• If you still want the algorithmic discovery of For You, run a sensationalism filter over it. Hide anything above 75; you'll lose almost nothing worth keeping

You will likely use X less after this setup. That is the point.

Don't bother with

A few things that look like they should help and don't:

• Following more "good" accounts: doesn't fix the For You tab, since FYP is mostly accounts you don't follow. Helps your Following tab, which you should be using anyway.
• Permanently logging out: works briefly, but then you've replaced X with not having X, which is a different question. Decide that one separately.
• "Just stop reading replies": replies are where rage-bait works hardest, and most people can't actually stick to this. Use the visibility filters on reply sections (Twitter Premium has "Hide replies below the threshold") if you can't.

The deeper point

X is not bad because of the posts. X is bad because the algorithm is optimised against the user. Filters fix specific symptoms; the only durable fix is changing the surface area you interact with — the Following tab, your list, your mutes. Treat the For You tab as something you visit on purpose, not the page you arrive at.