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

# Immersive Translate

Immersive Translate is a bilingual, side-by-side translation extension that uses large language models for cross-language understanding. It supports real-time translation of web pages, PDFs, EPUB e-books, and video subtitles, and is available as a browser extension and in other forms. Common use cases include:

* Read foreign-language articles or documentation with the original text and translation side by side.
* Translate PDF reports or EPUB books while keeping layout and structure.
* Generate or translate video subtitles for learning or accessibility.

You can use the ScitiX Model Inference API as the translation backend by configuring Immersive Translate to use the **OpenAI** or compatible API option and pointing it to the platform's Chat Completions endpoint.

## Configure

### 1. Install the extension

Open the [Immersive Translate](https://immersivetranslate.com/en/) official site and choose **Install browser extension** or the equivalent option. Pick the version for your browser, such as Chrome, Edge, or Firefox, and install it.

### 2. Configure API and model

Click the Immersive Translate extension icon and open **Settings**. In the settings page, set:

| Setting                 | Value                                                                                                    |
| ----------------------- | -------------------------------------------------------------------------------------------------------- |
| **API Key**             | Your API Key from the [API Keys](https://console.scitix.ai/model-inference/api_keys) page.               |
| **Model**               | The model's invocation name as shown on ScitiX, such as `Qwen/Qwen3.5-397B-A17B` or `DeepSeek-V4-Flash`. |
| **Custom API endpoint** | `https://api.scitix.ai/model-api/chat/completions`                                                       |

Save the settings.

**Tip**

* Keep your API Key private. Do not share it or commit it to version control.
* General-purpose or multilingual models, such as Qwen or DeepSeek models, work well for translation. Choose a model with sufficient context length for long documents.
* Translation of long texts can consume many tokens. Be aware of the platform's pricing and rate limits.

<img src="https://mintcdn.com/scitix-adfc65c2/zz3UuyM3SCvVbeP9/model-inference/media/immersive-translation-setup.png?fit=max&auto=format&n=zz3UuyM3SCvVbeP9&q=85&s=f9db80d63db7ee791f6976d49ad6dd44" alt="Immersive Translate — API and model configuration" width="2966" height="1520" data-path="model-inference/media/immersive-translation-setup.png" />

### 3. Set translation service to OpenAI

Set the **translation service** to **OpenAI**, or to the option that allows a custom API endpoint.

<img src="https://mintcdn.com/scitix-adfc65c2/zz3UuyM3SCvVbeP9/model-inference/media/select-api.png?fit=max&auto=format&n=zz3UuyM3SCvVbeP9&q=85&s=09a72c58446ed15360999eaa4486eab8" alt="Immersive Translate — select API / open settings" width="2984" height="1470" data-path="model-inference/media/select-api.png" />

## Verify

Click **Translate document** in the top-left to open the translation interface. Open a supported page, PDF, or EPUB, then use the extension to get bilingual translations.

The following example shows a page with the original text and translation side by side.

<img src="https://mintcdn.com/scitix-adfc65c2/zz3UuyM3SCvVbeP9/model-inference/media/sample-translation.png?fit=max&auto=format&n=zz3UuyM3SCvVbeP9&q=85&s=25834a75f541269a1c06ad40c6e5913e" alt="Immersive Translate — bilingual translation result" width="3000" height="1422" data-path="model-inference/media/sample-translation.png" />