> ## Documentation Index
> Fetch the complete documentation index at: https://docs.overlap.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Translate

> Turn clips into another language: translated captions, and optionally a dubbed voice.

> Deliver every clip in one language — even when the source mixes languages.

The **Translate** node creates a language version of each clip it receives. Captions are translated into the language you pick, and you can optionally replace the voice with a dubbed one. Each language version is its own clip in your library, tagged with the language, so you can render, edit, schedule, and post it like any other clip.

Use it when every clip must come out in one language. Without it, captions follow whatever language is spoken: a video that switches between English and Spanish gets English captions for the English parts and Spanish captions for the Spanish parts. With **Translate** set to Spanish, the whole clip is captioned in Spanish.

> [!NOTE]
>
> ### Schema
>
> * **Input**: Clips from an upstream node
> * **Output**: One translated clip per input clip, in the selected language. The original clips do not continue past this node.

## Add the node

Open **Workflows**, open your workflow, and add **Translate** after the nodes that style your clips (for example after [Style Video](https://docs.overlap.ai/nodes/style-video.md), Subtitles, or Title Overlay). The translated clip copies the styling of the clip it came from, so placing **Translate** last keeps your caption style, titles, and overlays.

> [!WARNING]
>
> Translate changes the language of existing captions; it does not add them.
> Add captions before it with [Style Video](https://docs.overlap.ai/nodes/style-video.md) (subtitles
> layer) or [Subtitles](https://docs.overlap.ai/nodes/subtitles.md).

## Configure it

* **Dubbing level**
  * **Subtitles only** — captions are translated; the original audio is kept.
  * **Subtitles + Voice** — captions are translated and the audio is replaced with a dubbed voice in the target language.
* **Target language** — one language per node: Spanish, French, German, Portuguese, Italian, Hindi, Japanese, Korean, Chinese, or Arabic. Regional variants such as Spanish (Mexico) or Portuguese (Brazil) are available on request.

## How it behaves

* Only the translated clips continue down the workflow. To also deliver the original, connect the upstream node to the rest of the workflow directly, alongside **Translate**.
* For several languages, add one **Translate** node per language, each connected to the same upstream node.
* Translated clips are named after the original with the language in parentheses, for example `Product Launch Highlights (Spanish)`, and carry a language tag.
* Only captions (and the voice, when dubbing) are translated. Title overlay text and the clip's title and description stay in their original language.
* Captions are translated line by line. Word-by-word highlighting is timed within each caption line rather than to each spoken word.
* If a clip cannot be translated, it is left out of the results rather than delivered in the wrong language.
* Translation usually takes one to a few minutes per clip, and clips are translated in parallel.

## Use it through the API

When a workflow contains a **Translate** node, API runs can set its language and dubbing level per request with `nodeConfigs.add_languages`. Runs that end in **Translate** return only the translated clips from `GET /workflow-results`. See [Translate Clips](https://docs.overlap.ai/api-reference/translate.md) for the full API flow.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.