SwearAtlas

About

Documenting language, not endorsing it

Colourful language. Worldwide.

Swear Atlas started with a very ordinary problem: someone says something in another language that sounds like koorva, and you have no idea what you just heard. Type what you heard — spelling entirely optional — and the Atlas works out the most likely word and explains what it actually means, not just its dictionary translation.

What we believe

Swearing is funny, human and linguistically fascinating. It's also a genuine part of understanding a culture. We treat ordinary profanity with curiosity and a wink.

Punchlines are welcome. Punching down isn't.

There's a real difference between ordinary profanity and language whose whole purpose is to attack people for who they are. Genuinely harmful slurs may still be documented — because someone may need to understand something that was said to them — but they're handled soberly. They never appear in suggestions, browsing, recommendations, trending lists, guides, the homepage spotlight, or on merchandise. They're only returned when a deliberate search strongly matches the word itself — including a close phonetic or transliterated spelling of it — and even then you get a single, plain identification, nothing more.

How search works

Most searches never need anything clever — the Atlas already knows the word. When you type a rough, heard-it-once spelling, it tries exact matches, known spellings, romanisation, accent-insensitive matching, edit-distance similarity and a light phonetic model, in that order. Only when the local catalogue genuinely can't help does it (optionally) ask an AI for a tentative, clearly-labelled suggestion — which never becomes a published page without independent validation.

Data & attribution

The full catalogue holds 1,484 canonical entries across 26 languages. Of these, 1,330 are freely browsable and searchable; the remainder are sensitive terms (slurs and similar) that we document for identification only — they don't appear in browsing, suggestions or search indexes, and are only returned when a deliberate search strongly matches the term itself, including a sufficiently close phonetic or transliterated spelling. Each term is attested against open-source word lists; the explanations, notes, Burn Ratings and pronunciation hints are original Swear Atlas editorial content. It is a source-attested, AI-curated reference, not a native-speaker-certified dictionary — meanings and severity are approximate and vary by context and region.

Swear Atlas uses these openly-licensed word lists as sources of attested vocabulary — evidence that a term exists. We select, normalise and map candidate terms, then write our own definitions, translations, pronunciation hints, aliases, classifications and Burn Ratings. Source material has been adapted; no third-party definitions are reproduced. No source maintainer endorses Swear Atlas.

MIT copyright and permission notices, and the CC BY 4.0 licence, are preserved with the dataset. Original Swear Atlas editorial content is © 2026 Swear Atlas. A complete source and licence register is kept in the project's THIRD_PARTY_DATA_LICENSES.md.

Merch

Some ordinary, good-natured entries are available as Swear Atlas stamp artwork on a mug or T-shirt. We don't run the shop ourselves — you complete your order with our print partner, Zazzle, who handle payment, printing, delivery, returns and support. Slurs, hate terms and threats are never eligible.

18+

This is an adults-only reference and contains explicit language throughout, by design.

Contact

Questions, corrections or content concerns? Email hello@swearatlas.com — one address for everything.