ko-KR · 한국어

Korean AI Voices

Korean is on the roadmap and not yet shipped. The hard part is not the voices; it is that speech level in Korean carries the relationship between characters, which a casting layer has to model rather than ignore.

Not available yet. There are no Korean voices in this release. This page exists to say what is planned and what it has to get right, not to sell you something that does not exist.

Speech levels are character information

Korean encodes the relationship between speakers in the verb ending of almost every sentence. Two characters speaking to each other tell a listener their relative age, closeness and social position before any content is considered. A tool that treats this as a rendering detail will produce dialogue that is grammatically fine and socially incoherent.

This means speech level has to live with the character relationship, not with the voice. It is a genuine design problem and it is why Korean is not simply a matter of adding voices to an existing pipeline.

Why Korean is on the list

Web novels, audio drama adaptations and vertical short drama are all large, active formats in Korean, and the production volume in the short-drama format specifically is very high. The demand for a multi-voice casting tool there is obvious.

It also matters to our English-language users: a large share of short-drama and webtoon-adjacent work is Korean in origin and localised outward, and supporting the source language properly is part of serving that pipeline.

What to do in the meantime

  1. Produce in English now

    Complete library, subtitles for reach.

  2. One line per utterance

    Keeps a later Korean track a drop-in rather than a restructure.

  3. Note the relationships

    If a Korean version is planned, record how characters address each other — that is the information a translator needs first.

  4. Plan for human translation

    Speech level decisions are authorial. They should not be automated.

  5. Tell us your use case

    Short drama, web novel adaptation and fan audio have different needs; which one you are is useful to know.

Frequently asked questions

Why is Korean harder than adding voices?

Because speech level is part of characterisation. Getting it wrong does not sound accented, it sounds like the characters have a different relationship than the story says they do.

When is it coming?

In the first group after English. No date until it is real.

Can I generate Korean now?

No. English voices reading Korean text produce nonsense, and we would rather say so than let you discover it.

Will honorific handling be automatic?

Partly. The plan is that the character card carries the relationship and the speech level follows from it, with manual override, rather than being guessed per line.

Cast your first scene tonight

Paste a script, let CastDub assign a voice to every character, and export a finished drama.

Create for free

No credit card. Free plan renews every month.