Wordlist

The default list covers Arabic script, Franco-arabe / Maghrebi Latin, French, English, and Tamazight (Kabyle, Chaoui, Mzabi, Saharan). Counts are exported as WORDLIST_STATS.

import { WORDLIST_STATS } from "anti-ttyah";

WORDLIST_STATS.terms
WORDLIST_STATS.phrases
WORDLIST_STATS.byLang      // ar, ar-latn, fr, en, tz
WORDLIST_STATS.byCategory  // sexual, insult, slur, mild

Do not paste the list into issues, READMEs, or social posts. Point at a file and a line range instead.

Extend without forking

Project-specific terms belong in createFilter, not in a pull request for every local insult:

const filter = createFilter({
  extraTerms: ["customslur"],
  extraPhrases: ["not allowed"],
  allowlist: ["h*ar"],
});

PROFANITY_TERMS and PROFANITY_PHRASES remain string[] for compatibility. Tagged entries live in PROFANITY_TERM_ENTRIES / PROFANITY_PHRASE_ENTRIES (term, lang, category).

Adding terms (contributors)

• Single tokens → PROFANITY_TERM_ENTRIES.
• Multi-word expressions → PROFANITY_PHRASE_ENTRIES.
• Do not add every Franco-arabe spelling. Normalization already folds 3/7/9, repeats, ة/ه, أ/إ/آ, and separators.
• Stay conservative. If a word is also a normal product name or given name, document an allowlist example instead.
• Add a focused test when matching behavior changes. Do not dump the whole list into a test file.