Internet Slang Decoder – What Slang Terms Mean, and Who Says So
Paste a message, comment or DM into this internet slang decoder and it marks the slang and abbreviations it recognises, showing a plain-English gloss for each one. You can also look up a single term. Everything runs in your browser against a glossary that ships with the page: nothing you paste is uploaded, and no meaning is produced on the fly.
Why slang meanings are contested
Slang is not issued by an authority. A term spreads through a community, picks up a second reading somewhere else, and drifts far enough that two people can use the same word to mean different things in the same thread. OP is the original poster in a forum and overpowered in a games chat. based once described someone addicted to crack cocaine and is now used approvingly of someone who ignores what others think. Because these readings coexist, this tool lists every sense it holds for a term rather than quietly picking one, and flags entries where sources materially disagree.
Why every entry carries a date
An undated slang list becomes wrong without ever looking wrong. Each entry here records the day it was checked and names the reference works it was checked against, and the glossary as a whole carries the same date on screen. That way a reading you find here can be judged as what it is: a snapshot of informal usage, taken on a particular day, not a permanent definition. The dataset is also revisited on a schedule, because new terms appear continuously and old ones shift.
Why “not in our glossary” is the honest answer
A decoder could expand any string of capitals into something plausible. That is exactly what this one refuses to do. Lookup is an exact match against stored forms — there is no acronym expander, no pattern that turns unfamiliar letters into a definition, and no model generating text at run time. When a single term has no exact match, the tool offers spellings that genuinely exist in the dataset, and when nothing is close it offers nothing at all.
How the text is read
Your text is split with Intl.Segmenter at a pinned locale, so words are found the same way in every browser that has it, and matching folds case in that same locale rather than the viewer’s — folding IYKYK under a Turkish locale would produce a dotless ı and miss the entry entirely. Text is normalised to NFC first so composed and decomposed spellings agree. Phrases of up to five words are tried before single words, so spill the tea wins over tea, and matched spans never overlap.
Hashtags and @handles are read as entities, not words, so #iykyk is not looked up. Web and email addresses are skipped for the same reason, and emoji stay whole instead of being shredded into stray tokens. Wherever a length is shown, three distinct numbers are labelled separately: graphemes (characters a reader sees), code points (Unicode scalar values), and UTF-16 code units (what String.length returns). None of them is presented as “characters” without saying which.
What is deliberately left out
Some internet slang is sexual, and some is a slur. Slurs and hate terms are not in this dataset at all. Entries whose sense is sexual are marked sensitive, kept out of the inline reading, and collapsed behind a click, with a switch that hides them by default. This tool describes what terms are used to mean; it does not tell you how to write, and it makes no claim that using or avoiding any term changes how a post travels.