Fantasy NamesThe name table

The honest summary

The names on this site are generated by AI models and checked by machines. No human linguist has reviewed the whole dataset yet. Errors may remain, and some meanings, readings and era notes will be wrong or debatable. Treat every entry as a careful suggestion, not as a reference book. If you need a name for a real person, a real brand or a published translation, check it with a native speaker.

How an entry is made

  1. AI models propose names with a meaning, origin, era and respelling.
  2. A second AI pass with web search checks the facts and corrects or rejects entries.
  3. Code checks compare Korean hanja readings, Chinese characters and Japanese kanji with the Unicode Unihan database, KANJIDIC2 and JMnedict. Korean Revised Romanization is computed from the Hangul by code.
  4. Names are screened against a hand-made list of famous characters and trademarks. Matches that are plainly protected are removed, and weaker links get an information note. The list is not complete.
  5. Names are screened against lists of vulgar or unlucky words in English, Spanish, French, Italian, German, Tagalog, Indonesian, Japanese, Korean and Chinese. An AI review kept the meaningful warnings and dropped trivial ones. Slang in other languages, such as Hindi, Thai, Malay and Arabic, is not systematically checked.
  6. Entries with low confidence are not published.

Audits and what they found

Random samples are re-checked by a third, skeptical AI pass with web search. In the last audit of 217 random entries, 91.7% were fully correct and 1.4% were wrong; the rest had minor issues. The two earlier audits found 90.8% and 89.4% accurate, and 3.2% and 3.7% wrong. So a small share of entries is likely still wrong. These audits are done by AI too, so they are an estimate, not proof.

Labels you will see

High confidence means a second AI check with web search found support and the code checks passed. Medium means the facts were corrected or hedged, or could not be fully matched in the reference databases. Origin debated means sources disagree about where the name comes from.

Overused flags come from complaints and discussions of readers on Reddit about names that appear in many stories, plus a few editorial picks. They are not a measured count. The fresher alternatives are chosen by sound and style.

Check meaning warns that the name or a syllable may sound like a rude, silly or unlucky word in some language.

Known limits

Licence line and corrections

Character data: Unicode Unihan Database; KANJIDIC2 and JMnedict © EDRDG (CC BY-SA 4.0).

To report an error, email [email protected] with the name, the page, and what is wrong. A source link helps. See also the contact page.

Questions

Is any of this verified by experts?

Not fully. Checks are done by AI and by code. A human review of the whole dataset has not happened yet.

Why remove some famous names?

Names tied to a well-known character or trademark are left out so you do not copy a protected name by accident. The list is hand-made and not complete.