# How do you add music to an AI drama?

**Short answer:** Score in the edit, not in the video model: generate or license music separately, place it under the episode and mix to a consistent loudness. We mix finished episodes to about -14 LUFS, while raw AI renders came out around -17 to -19 LUFS.

URL: https://arc.katechon.technology/answers/how-to-add-music-to-an-ai-drama
Updated: 2026-10-18

## Keep music out of the shot prompt

Video models with native audio can add music, but you cannot control it across cuts: each clip picks its own key, tempo and volume, and a cut between two clips becomes a cut between two songs. We ask for dialogue and ambience in the shot and add music afterwards over the whole scene.

## What we used

Music in our own productions:

| Production | Music | How it was used |
| --- | --- | --- |
| [High Note](https://arc.katechon.technology/films/high-note) | 9 original songs made with Suno | Mixed into the episodes as story moments; the series is about a singer |
| The Little Fairy Bakery | Eight music set pieces generated with ElevenLabs music | Cued under the music scenes of 30 vertical episodes |
| [Dragon Rising](https://arc.katechon.technology/films/dragon-rising) | A generated score with six chapter cues plus a sound-effects pack | Gated in the edit so silence lands before each hit |
| [Blood Sun: The Fall of Kratos](https://arc.katechon.technology/films/blood-sun) | No music | Shipped as rendered; one of the things we would change |

## Loudness

Raw AI clips are quieter than platform norms. Several of our first series shipped at -17 to -19 LUFS with no music, noticeably quieter than the videos around them in a feed. Our later episodes are mixed to about -14 LUFS integrated, so episodes sit at a similar level to the videos around them.

## Steps

1. Lock the edit with dialogue and ambience only.
2. Choose cues per scene, not per clip: one cue should run across several cuts.
3. Duck music under dialogue so every line stays intelligible.
4. Leave silence before big moments; a hit lands harder after a gap.
5. Normalize the finished episode to about -14 LUFS and check the loudest moment does not clip.
6. Check the music rights for the platform you publish to; generated music has its own terms.

## Where Katechon Arc fits

Arc gives you each episode's dialogue and ambience, plus a Jianying draft export so the music pass happens in an editor on top of the assembled cut.

Arc does not compose music or mix and normalize loudness for you; that is an editor's job after export.

## Related answers

- [How do you make an AI music video?](https://arc.katechon.technology/answers/how-to-make-an-ai-music-video): Start from a finished song, plan shots to its sections, keep the performer consistent with a reference sheet, generate clips, and cut them…
- [How do you add voice and dialogue to AI video?](https://arc.katechon.technology/answers/how-to-add-voice-and-dialogue-to-ai-video): Let the video model speak the lines that are on screen and add narration separately with text-to-speech.
- [How do you add accurate captions to AI video?](https://arc.katechon.technology/answers/how-to-add-captions-to-ai-video): Transcribe the final mix with speech recognition, then correct it against the script, especially invented names.
- [How do you make a movie trailer with AI?](https://arc.katechon.technology/answers/how-to-make-an-ai-trailer): Lock the look and the cast first, build every location as a set of camera angles, write each shot as one action with a concrete light…
- [How do you make AI video look less like AI?](https://arc.katechon.technology/answers/how-to-make-ai-video-look-less-ai): Pick one visual language and hold it everywhere, keep faces and places consistent with references, and put colour and lighting words on the…