# How do you add accurate captions to AI video?

**Short answer:** Transcribe the final mix with speech recognition, then correct it against the script, especially invented names. On Blood Sun: The Fall of Kratos, aligning the script to speech-recognition word timings fixed 396 of 400 caption lines that planned timings had placed early.

URL: https://arc.katechon.technology/answers/how-to-add-captions-to-ai-video
Updated: 2026-10-18

## Why captions from the script drift

A script says what each shot should contain, not when each word lands. Generated clips speak at their own pace, trim differently in the edit and sometimes add or drop a word. On [Blood Sun: The Fall of Kratos](https://arc.katechon.technology/films/blood-sun), captions timed from the plan could sit a whole line early. Re-timing them against speech-recognition word timings fixed 396 of 400 lines.

## Why speech recognition alone is not enough

Speech-to-text gets common words right and invented names wrong. The first caption track of [The Fall of Kalak](https://arc.katechon.technology/films/the-fall-of-kalak) had six name errors. Fantasy and drama scripts are full of names, so check every proper noun before upload.

## The method we use

1. Lock the edit and mix the final audio.
2. Run speech recognition on the mix to get word-level timings.
3. Align the script text to those timings, so you keep the script's spelling and the audio's timing.
4. Keep the speech-recognition wording where the generated clip changed a word, so captions match what is heard.
5. Check proper nouns, then check any line the model invented (caption it or cut it).
6. Export as a separate subtitle track for long films; burn captions in for vertical episodes, which are often watched without sound.

## Burned in or separate?


| Format | Captions | Why |
| --- | --- | --- |
| Vertical episodes (9:16) | Burned in | Feeds autoplay muted; burned-in English captions also told one platform's language detection that the series was in English |
| Long 16:9 films | Separate subtitle track | Viewers can turn them off; we published The Wolf Mother and The Fall of Kalak with English caption tracks |
| Trailers and clips | Burned in | Short, often muted, reposted out of context |

## Where Katechon Arc fits

Arc keeps the exact script text for every shot, which is what you align recognized speech against, and exports episodes and a Jianying draft for captioning in an editor.

Arc does not transcribe or time captions for you; captioning happens after the edit, in your editor or a captioning tool.

## People also ask

### Should captions be burned into vertical AI dramas?

Yes, for vertical episodes. They autoplay muted in feeds, and readable captions keep viewers who have sound off.

## Related answers

- [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.
- [Why does AI video add voices or lines you did not write?](https://arc.katechon.technology/answers/why-does-ai-video-add-random-voices): Native-audio video models fill silence: given a near-empty shot, they sometimes invent a voice-over or a stray line.
- [How do you add music to an AI drama?](https://arc.katechon.technology/answers/how-to-add-music-to-an-ai-drama): Score in the edit, not in the video model: generate or license music separately, place it under the episode and mix to a consistent…
- [Should an AI drama be vertical (9:16) or horizontal (16:9)?](https://arc.katechon.technology/answers/vertical-or-horizontal-for-ai-drama): Vertical 9:16 for short dramas watched in phone feeds and short-drama apps; horizontal 16:9 for films and long-form YouTube.
- [What goes wrong when you make an AI series?](https://arc.katechon.technology/answers/what-goes-wrong-when-making-an-ai-series): Mostly waste and drift, not catastrophe: across our first five series 15% of video spend went to re-renders, we logged about 50 hours of…