Muted words that catch bot replies on X (a starter list, and why it's not enough)

By Jose Enrique, maker of BotBlind · Updated October 19, 2026 · 7 min read

How to add a muted word

1. Open Settings in X.
2. Find Muted words. It sits in the privacy and safety part of Settings, next to your muted and blocked accounts. At the time of writing the path on the web is Settings and privacy → Privacy and safety → Mute and block → Muted words. X renames menus from time to time, so if you don't see it, search Settings for "muted".
3. Add a new entry and type one word or phrase. X takes one entry at a time.
4. Choose where it applies (your Home timeline, your notifications, or both), from whom (anyone, or only people you don't follow) and for how long (forever, or a limited time such as a day, a week or a month). Save.

A few things worth knowing before you start. Muting is matched against the text of the post, so a phrase inside an image won't be caught. Muting a word also mutes it as a hashtag. And muted words don't apply to search results (Botfilter summarises X's help pages on this).

Menu names and option labels change over time. If a label here doesn't match what you see, the setting is still there under a similar name.

The settings that keep real people visible

Most of the damage muted words do is to people you actually wanted to hear from. Three habits avoid most of it:

• Limit risky mutes to people you don't follow for anything that a friend might plausibly write. Bots replying to big posts are almost never people you follow.
• Mute phrases, not words. check my bio is a spam hook. bio is a word your friends use about their biology exam.
• Use a time limit (a month is a good start) for anything risky, then look at what you missed. If it was only spam, add it again with no time limit.

The starter list

Sorted by how likely each phrase is to catch a real person. These come from the spam hooks and stock phrases BotBlind's rules look for. Treat it as a starting point and trim anything that doesn't fit your corner of X.

Tier 1: spam hooks (low risk: from anyone, forever)

These are calls to action that pull you off the thread. Real people rarely write them in replies to strangers.

• check my bio
• check my pinned
• look at my bio
• link in bio
• dm for promo
• promo available
• dm me for
• message me on whatsapp
• claim your airdrop
• free airdrop
• 100x gem
• giveaway is live

Tier 2: AI-style praise (medium risk: only people you don't follow, for a month)

The stock lines that fill reply sections under big posts. People write these too, which is why they get the safer settings.

• great point
• couldn't agree more
• this resonates
• a testament to
• game-changer and game changer (add both)
• thanks for sharing this insight
• this is gold
• well said
• absolutely spot on
• so true

Tier 3: don't mute these

You'll see them suggested in "mute these to fix X" posts. They cost more than they save.

• Single words like delve, insightful or amazing. Too many real posts use them, not only replies.
• Emojis like 🚀 or 🔥. You'll lose half of sports and launch-day X.
• Punctuation like the em dash. Many careful writers use it.
• Topics you're curious about just because bots like them (AI, crypto). Mute the hook, not the subject.

Muted words not working? Check these

1. Where it applies. Edit the entry: if it only applies to notifications, it won't touch your timeline.
2. From whom. If it only applies to people you don't follow, posts from accounts you follow still show. That's working as set.
3. Expiry. Time-limited mutes disappear quietly when they run out. If an entry is gone from the list, it expired.
4. Search. Muted words don't apply to search results.
5. Spelling. Spammers write ch3ck my b1o, use lookalike letters, or split words with dots. You'd need an entry for each variant, which is where lists stop scaling.
6. Text in images and links. A screenshot that says "check my bio" isn't text, so it can't be muted.
7. Web vs app. If the web hides something and the app doesn't, update the app and check the same entry there.

Why a list is never enough

Muted words are a filter by wording. Bot replies change wording constantly, and the AI-written ones barely repeat at all. So a list catches the lazy spam and misses the rest, while slowly picking up false positives as you add broader phrases. It also needs upkeep: every new template means another entry, typed one at a time.

That's fine. Use the list for what it's good at, the spam hooks, and use other signals for the rest: timing, copies across the thread, and the account's history. Our guide on how to tell if a reply on X is from a bot walks through those, and how to hide AI and bot replies on X covers every other native setting.

Or let something read the whole thread

BotBlind is a free Chrome extension that does what muted words can't: it looks at each reply in context. A phrase like "great point" adds a few points, but so does landing eight seconds after the post, being copied by three other accounts, or sharing no words with the post. Replies over your threshold fold into one line with the reasons, and Show brings any of them back. It runs entirely in your browser, needs no X login and sends nothing anywhere.

It works well alongside a short Tier 1 list. Keep the hooks muted, and let the scoring handle the rest.

Status: BotBlind's Chrome Web Store listing is pending.

FAQ

Do muted words hide replies for everyone?

No. Muting only changes what you see. The reply stays on X, other people still see it, and the author isn't told.

Why do I still see posts with my muted words?

The most common reasons are that the mute only applies to notifications, it only applies to people you don't follow, it has expired, you are looking at search results, or the post spells the phrase differently. Edit the entry and check each option.

Can muted words stop AI-written replies?

Only the ones that reuse a phrase you listed. AI replies rarely repeat the exact same wording, and the phrases they do reuse are also used by real people. Muted words work best against spam hooks like "check my bio", not against AI writing in general.