What is algospeak?
Algospeak is a way of talking around filters. People replace words they believe automated moderation will flag with look-alike spellings and terms. The word comes from "algorithm" plus "speak." It grew on short video apps, where creators worried that certain words would limit their reach or cut their earnings.
A well-known example is "unalive" in place of words about death or suicide. Others include "seggs" for sex and "le dollar bean" for lesbian, which came from a text filter turning the word into odd symbols. Some terms fade within months. Others spread so far that people use them in everyday speech, far from any filter.
Nobody official decides what counts as algospeak. Terms start in one corner of an app, get copied in comments and spread when a big creator picks them up.
How algospeak works
Most algospeak follows a handful of patterns. Once you know them, new terms are easier to spot.
- Swapped letters. Numbers or symbols stand in for letters, such as "s3x" or "k!ll."
- Stand-in words. A cute or harmless word replaces the real one, such as "unalive" or "spicy accountant."
- Spoken one way, written another. A creator says the real word out loud while the captions or on-screen text show a softer version.
- Emoji codes. A single emoji carries a meaning that insiders know, such as a corn emoji used for "porn" in some circles.
- Sound-alikes. A word that sounds close, such as "seggs," lets the speaker keep the meaning while dodging exact matches.
Where people use algospeak
You will see it most on TikTok, Instagram Reels and YouTube Shorts, where much of the reach comes from recommendations and creators feel the most pressure to stay "safe." It shows up in captions, on-screen text, hashtags and comments. Creators who talk about health, news, money or hard personal topics use it most, because those subjects are more likely to be reviewed.
Text-heavy networks see it less often, though it still appears in comment sections and replies. On open networks such as Mastodon, people more often use content warnings, which hide a post behind a short label until the reader taps to open it.
An example of algospeak
Here is an example post: "That movie scene where the dog gets unalived had me crying for an hour #movietok." The writer chose "unalived" to avoid a word they feared would push the video down. A viewer who knows algospeak reads it right away. Someone new to the app might be confused. A screen reader would read it out as written, which may make little sense to a listener.
| Plain words | Algospeak | |
|---|---|---|
| Easy to find in search | – | |
| Clear to new readers | – | |
| Reads well on a screen reader | – | |
| Understood by insiders | ||
| Keeps a post within the rules | Depends on the content | Depends on the content |
Why algospeak matters for brands
Even if you never use it, algospeak changes how people talk about topics near your brand. That affects how you listen and how you moderate.
- Social listening. If you track mentions of a topic, you may miss posts that use coded words. Add common variants to your keyword lists and review them every few months.
- Moderation. Comment filters that block exact words can miss algospeak. Review hidden and flagged comments by hand from time to time.
- Clarity. Coded words can confuse new readers, older audiences and people using screen readers.
- Tone. Some algospeak makes serious topics sound light. A health brand that copies "unalive" may come across as careless.
- 1List your topics
Write down subjects near your brand that people may talk about in code.
- 2Collect variants
Note the spellings, stand-ins and emoji people use for each one.
- 3Update your filters
Add the variants to listening keywords and comment filters.
- 4Review by hand
Check hidden and flagged comments every week or two.
Should you use algospeak?
For most brands, the answer is usually no. Platforms publish community guidelines that explain what they limit, and many sensitive topics are allowed when handled with care. Odd spellings can make your post harder to find in search and harder to understand. If a topic is truly against the rules, coded words won't make it safe, and moderation systems learn new terms over time.
Write clearly and follow the guidelines. When you cover hard subjects, add context, use a content warning where the network offers one and link to support resources. If you work with creators who use algospeak in their own voice, agree up front on which words appear in posts made for your brand.
Common mistakes
- Using algospeak in a brand post to sound current, then confusing most of your audience
- Building comment filters around exact words and missing the coded versions
- Assuming every odd spelling is harmless, when some codes hide harassment or scams
- Copying a term without knowing what it means in the community that made it
Related terms
- AlgorithmAn algorithm is the set of rules a social platform uses to decide which posts to show each person and in what order.
- ModerationModeration is reviewing and managing the comments and messages in your community to remove harm and keep talk useful.
- Community guidelinesCommunity guidelines are a platform's or group's public rules for what people may post, and what happens when someone breaks them.
- ShadowbanA shadowban is an informal name for when an app limits who sees an account's posts without telling the owner, so reach drops with no clear reason.
- Social listeningSocial listening is tracking what people say online about your brand, competitors and industry, then using what you learn to make better decisions.