Language Name Generator

How this list was compiled: Opening syllables written for this page, stored with a capital letter because each one starts the name. The tool builds a language name from three rows (opening, stem end, real ending) and adds a fourth row that says who speaks it. Rules for this row: every line ends in a vowel, so any stem end can follow it; and no line may spell, once a stem end is added, the front of a real people, country or faith, or a name from a published game, novel, film or invented language. Ira, Oma, Ura, Ava, Atha, Da, Sve, Ke, Che, Gre, Bre, Ne, Ve, Ru, Mo, Ta, Ty, Zu and Kla were written and then removed because they spelled real peoples or places. The opening consonant and its vowel share one row on purpose: with a separate vowel row real words and real peoples could not be kept out. The tool cannot block one combination, so every name the three name rows can spell is built and screened. The first screen, on 2026-09-19, was for real languages, peoples and places: the ISO 639-3 code table and name index published by the registrar, SIL International (https://iso639-3.sil.org/sites/iso639-3/files/downloads/iso-639-3.tab and iso-639-3_Name_Index.tab in the same folder); the Glottolog languoid list (https://raw.githubusercontent.com/glottolog/glottolog-cldf/master/cldf/languages.csv); Wikipedia's lists of demonyms for countries and for subcontinental regions, of constructed languages, of language names in their own languages, of country names in native languages, of national capitals, largest cities, sovereign states, Indigenous peoples and ancient peoples of Italy (https://en.wikipedia.org/wiki/List_of_adjectival_and_demonymic_forms_for_countries_and_nations, https://en.wikipedia.org/wiki/List_of_adjectivals_and_demonyms_for_subcontinental_regions, https://en.wikipedia.org/wiki/List_of_constructed_languages, https://en.wikipedia.org/wiki/List_of_language_names, https://en.wikipedia.org/wiki/List_of_countries_and_dependencies_and_their_capitals_in_native_languages, https://en.wikipedia.org/wiki/List_of_national_capitals, https://en.wikipedia.org/wiki/List_of_largest_cities, https://en.wikipedia.org/wiki/List_of_sovereign_states, https://en.wikipedia.org/wiki/List_of_Indigenous_peoples, https://en.wikipedia.org/wiki/List_of_ancient_peoples_of_Italy); the macOS system dictionary of English words and proper names; a 50,000-word English frequency list (https://github.com/hermitdave/FrequencyWords); a hand list of peoples and languages from well-known fiction; and a list of rude words. A review the same day showed that screen was far too thin on fiction: the rows could then spell Daedric, Lothric, Redoran, Revali and Azazel, and stems such as Ryloth and Lolth, all names from published games, films or older lore. So a second screen was run the same day over every stem (an opening plus a stem end) and every full name, against: the text of 232 English Wikipedia pages, nearly all of them articles and lists on fictional worlds, peoples, gods and invented languages (among them https://en.wikipedia.org/wiki/List_of_fictional_countries_set_on_Earth, https://en.wikipedia.org/wiki/List_of_Star_Wars_planets_and_moons, https://en.wikipedia.org/wiki/List_of_Dungeons_%26_Dragons_deities, https://en.wikipedia.org/wiki/List_of_Cthulhu_Mythos_deities, https://en.wikipedia.org/wiki/The_Elder_Scrolls, https://en.wikipedia.org/wiki/Dark_Souls_III and the 26 lettered pages under https://en.wikipedia.org/wiki/Lists_of_fictional_extraterrestrial_species_and_races); the 5,177 page titles in 29 Wikipedia categories such as Category:Fictional_languages and Category:Fictional_countries; about 1.2 million page titles from 39 fan wikis for well-known games, novels, films and shows, read through each wiki's public page index (seven of the largest were read only part of the way through the alphabet); and the GeoNames lists of towns and cities of more than 500 people (https://download.geonames.org/export/dump/). A stem or a name counted as a hit when it matched a word from those sources that is not an English dictionary word, or when a well-known fictional name of six letters or more sat inside it. The rows as first written produced 1,241 hits, and 32 openings and 24 stem ends were removed. A second pass with more Wikipedia pages found 197 more hits in what was left (Zenn-Lavian, Azathi, Khoros and the Naʼvi word leNaʼvi), so one more opening and three more stem ends came out, and two new stem ends, zith and vuth, went in after passing both screens. Openings removed: La, Le, Lo, No, Ra, Re, Ro, Va, Vo, Za, Zo, The, Te, To, Je, Khe, Kho, Dra, Dre, Dro, Gra, Kra, Kre, Tra, Rau, Tau, Ala, Alo, Ela, Eli, Ori, Uma, Otha. Stem ends removed: zel, zun, val, niv, lav, nav, lth, thr, thel, thur, reth, nath, zath, loth, rith, dr, ldr, dor, vok, lder, lden, lusk, vask, lend, veld, neld, nirn, and in the third check kol, kr and lun. That leaves 59 openings x 50 stem ends x 7 endings = 20,650 names, and none of them matches anything in the lists above. A third check on 2026-09-20 sent every name and every stem as an exact title to the English Wikipedia API (https://en.wikipedia.org/w/api.php, redirects followed), which holds what the lists above do not: Kaikolan redirects to the article on the Sengunthar, a caste community of Tamil Nadu and Sri Lanka; Tzakol is a creator god in Maya tradition; Thikri, Thikrian and Myladi are towns in India and Pakistan too small for the GeoNames list; Naukri is a 1978 film; and Kylun is a Marvel Comics character. The stem ends kol, kr and lun came out for that, and tev, zev and nuv went in after passing both earlier screens, the same Wikipedia title check (every name and every stem they spell) and an exact title check on 21 fan wikis (one more, awoiaf.westeros.org, refused the requests). The count is unchanged at 59 x 50 x 7 = 20,650. 21 stems with no ending are still Wikipedia titles for a small village, a surname or a given name (Rynek, Reidun, Olakur); none names a people, a faith or a language, and they are listed in the held file. That is a screen and not a promise. Seven fan wikis were not read to the end, no list of fiction is complete, and a short invented word can always turn out to be somebody's character or village. Search a name before you publish it. Four groups by sound: soft, clipped, long vowels and vowel first. Licence: own work

How this list was compiled: Stem ends written for this page, in lower case because each one sits in the middle of the name, between the opening and the real ending. Rules for this row: every line starts and ends with a consonant, so it can follow any opening and carry any of the seven endings (after vr you get Lyvrish, Lyvrian and Lyvri); no line ends in a letter pair that doubles with an ending; and no line builds a real name with an ending, which is why lev (Levite), lind (Indian), thol (Catholic), rev, riv, dorn, rmen, str, nsk and mir were written and then removed. The tool cannot block one combination, so every name the three name rows can spell is built and screened. The first screen, on 2026-09-19, was for real languages, peoples and places: the ISO 639-3 code table and name index published by the registrar, SIL International (https://iso639-3.sil.org/sites/iso639-3/files/downloads/iso-639-3.tab and iso-639-3_Name_Index.tab in the same folder); the Glottolog languoid list (https://raw.githubusercontent.com/glottolog/glottolog-cldf/master/cldf/languages.csv); Wikipedia's lists of demonyms for countries and for subcontinental regions, of constructed languages, of language names in their own languages, of country names in native languages, of national capitals, largest cities, sovereign states, Indigenous peoples and ancient peoples of Italy (https://en.wikipedia.org/wiki/List_of_adjectival_and_demonymic_forms_for_countries_and_nations, https://en.wikipedia.org/wiki/List_of_adjectivals_and_demonyms_for_subcontinental_regions, https://en.wikipedia.org/wiki/List_of_constructed_languages, https://en.wikipedia.org/wiki/List_of_language_names, https://en.wikipedia.org/wiki/List_of_countries_and_dependencies_and_their_capitals_in_native_languages, https://en.wikipedia.org/wiki/List_of_national_capitals, https://en.wikipedia.org/wiki/List_of_largest_cities, https://en.wikipedia.org/wiki/List_of_sovereign_states, https://en.wikipedia.org/wiki/List_of_Indigenous_peoples, https://en.wikipedia.org/wiki/List_of_ancient_peoples_of_Italy); the macOS system dictionary of English words and proper names; a 50,000-word English frequency list (https://github.com/hermitdave/FrequencyWords); a hand list of peoples and languages from well-known fiction; and a list of rude words. A review the same day showed that screen was far too thin on fiction: the rows could then spell Daedric, Lothric, Redoran, Revali and Azazel, and stems such as Ryloth and Lolth, all names from published games, films or older lore. So a second screen was run the same day over every stem (an opening plus a stem end) and every full name, against: the text of 232 English Wikipedia pages, nearly all of them articles and lists on fictional worlds, peoples, gods and invented languages (among them https://en.wikipedia.org/wiki/List_of_fictional_countries_set_on_Earth, https://en.wikipedia.org/wiki/List_of_Star_Wars_planets_and_moons, https://en.wikipedia.org/wiki/List_of_Dungeons_%26_Dragons_deities, https://en.wikipedia.org/wiki/List_of_Cthulhu_Mythos_deities, https://en.wikipedia.org/wiki/The_Elder_Scrolls, https://en.wikipedia.org/wiki/Dark_Souls_III and the 26 lettered pages under https://en.wikipedia.org/wiki/Lists_of_fictional_extraterrestrial_species_and_races); the 5,177 page titles in 29 Wikipedia categories such as Category:Fictional_languages and Category:Fictional_countries; about 1.2 million page titles from 39 fan wikis for well-known games, novels, films and shows, read through each wiki's public page index (seven of the largest were read only part of the way through the alphabet); and the GeoNames lists of towns and cities of more than 500 people (https://download.geonames.org/export/dump/). A stem or a name counted as a hit when it matched a word from those sources that is not an English dictionary word, or when a well-known fictional name of six letters or more sat inside it. The rows as first written produced 1,241 hits, and 32 openings and 24 stem ends were removed. A second pass with more Wikipedia pages found 197 more hits in what was left (Zenn-Lavian, Azathi, Khoros and the Naʼvi word leNaʼvi), so one more opening and three more stem ends came out, and two new stem ends, zith and vuth, went in after passing both screens. Openings removed: La, Le, Lo, No, Ra, Re, Ro, Va, Vo, Za, Zo, The, Te, To, Je, Khe, Kho, Dra, Dre, Dro, Gra, Kra, Kre, Tra, Rau, Tau, Ala, Alo, Ela, Eli, Ori, Uma, Otha. Stem ends removed: zel, zun, val, niv, lav, nav, lth, thr, thel, thur, reth, nath, zath, loth, rith, dr, ldr, dor, vok, lder, lden, lusk, vask, lend, veld, neld, nirn, and in the third check kol, kr and lun. That leaves 59 openings x 50 stem ends x 7 endings = 20,650 names, and none of them matches anything in the lists above. A third check on 2026-09-20 sent every name and every stem as an exact title to the English Wikipedia API (https://en.wikipedia.org/w/api.php, redirects followed), which holds what the lists above do not: Kaikolan redirects to the article on the Sengunthar, a caste community of Tamil Nadu and Sri Lanka; Tzakol is a creator god in Maya tradition; Thikri, Thikrian and Myladi are towns in India and Pakistan too small for the GeoNames list; Naukri is a 1978 film; and Kylun is a Marvel Comics character. The stem ends kol, kr and lun came out for that, and tev, zev and nuv went in after passing both earlier screens, the same Wikipedia title check (every name and every stem they spell) and an exact title check on 21 fan wikis (one more, awoiaf.westeros.org, refused the requests). The count is unchanged at 59 x 50 x 7 = 20,650. 21 stems with no ending are still Wikipedia titles for a small village, a surname or a given name (Rynek, Reidun, Olakur); none names a people, a faith or a language, and they are listed in the held file. That is a screen and not a promise. Seven fan wikis were not read to the end, no list of fiction is complete, and a short invented word can always turn out to be somebody's character or village. Search a name before you publish it. Four groups by sound: soft, breathy, clipped and heavy. Licence: own work

How this list was compiled: The seven endings that English really uses to name a language or a people, in lower case because each one closes the name. Each ending's origin was checked on 2026-09-19 against its English Wiktionary entry (https://en.wiktionary.org/wiki/-ish, -ese, -ian, -an, -ic, -i and -ite on the same pattern; every URL is in the item's x.src), and -ish also against https://www.etymonline.com/word/-ish. The notes are in our own words and leave out the accents the dictionaries print. The example languages are reference names that end in those letters in the ISO 639-3 code table from SIL International, the standard's registrar (https://iso639-3.sil.org/sites/iso639-3/files/downloads/iso-639-3.tab, checked 2026-09-19); every one except Hittite and Elamite also has a two-letter code on https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes, checked the same day. Of the 184 names with a two-letter code, 30 end in -ian, 12 in -ish, 12 in -ese and 7 in -ic. None ends in -ite. The full table has nine reference names that end in those letters. Seven carry the ending and are typed as historical: Elamite, Hittite, Old Hittite, Middle Hittite, Neo-Hittite, Moabite and Edomite. The other two, Kanite and Nde-Gbite, are typed as living languages, and their names only happen to end in those letters, so they are not used as examples. Wiktionary gives -ite for peoples and followers and does not list a language sense, so the page says only that the table uses it for old tongues. Examples were picked where the ending is plainly added to a place or people, so German, Catalan, Thai and Somali are not used. Left out because they could not be verified as language endings: -ean, -ac and -aic. Two groups of Latin descent and borrowed endings, plus the one Germanic ending. Licence: own work

How this list was compiled: An original list written for this page: 60 lines that say who speaks the language and what it is for. Rules for this row: each line is a phrase that follows the name and a comma, starts in lower case and takes no full stop, because the tool adds it. No line names a people, a place on a real map, a real language or a real faith, and none uses a name or term from a published game, novel, film or invented language. No line depends on the name's ending, so all 60 read well after any of the 20,650 names. The thieves, smugglers and spies are storybook ones: no real group and no real method. Six groups of 10: trade tongues, court and law, temple and rite, hearth and home, guild and secret, fading and lost. Licence: own work