Overview

Project website app.immersivetranslate.com
Instructions for translators

Immersive Translate Document Translation
Project maintainers User avatar theowenyoung User avatar zhoulink90 User avatar marshmolly2b
3 days ago

String statistics

Strings percent Hosted strings Words percent Hosted words Characters percent Hosted characters
Total 2,525 26,000 182,344
Approved 0% 0 0% 0 0% 0
Waiting for review 93% 2,364 90% 23,441 89% 163,646
Translated 93% 2,364 90% 23,441 89% 163,646
Needs editing 2% 57 7% 1,897 7% 13,644
Read-only 0% 0 0% 0 0% 0
Failing checks 15% 385 21% 5,625 21% 38,762
Strings with suggestions 0% 0 0% 0 0% 0
Untranslated strings 4% 104 2% 662 2% 5,054
Last 12 months Previous

Activity in last 12 months

0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 0 changes 2,217 changes 0 changes 454 changes 0 changes 163 changes 0 changes 10 changes
May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr

Quick numbers

26,000
Hosted words
2,525
Hosted strings
93%
Translated
1
Contributors
and previous 30 days

Trends of last 30 days

+1%
Hosted words
+100%
+1%
Hosted strings
+100%
+1%
Translated
+93%
—
Contributors
+100%
User avatar None

Resource updated

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
The “app/src/i18n/zh-HK.json” file was changed. 13 days ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
We currently use a "translation quota" based billing method, detailed as follows:<br/><br/><ul><li>Translation quotas are charged based on the number of <b>Tokens</b> used during the translation process.</li><li>The <b>GLM-4-Flash-250414</b> model is available to all users, and every BabelDOC user receives a monthly quota of 500,000 Tokens for free model translations within BabelDOC. (Actual availability depends on the supply from upstream providers, and quotas are dynamically adjusted based on server load)</li><li>The <b>Qwen3.5-Plus/DeepSeek-V4-Flash/doubao-seed-2-0-lite/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7</b> models are exclusively for <b>Pro/Max</b> members, counted towards the members' general translation quota. Pro members enjoy up to <b>20 million Tokens</b> per month, while Max members enjoy up to <b>50 million Tokens</b> per month. Failed translations are not counted towards the quota.</li></ul><br/>At the start of a translation task, the system will deduct the general translation quota in advance based on "3000 Tokens × number of pages". Once the translation is complete, we will settle based on actual consumption, refunding any excess to your account or requesting additional tokens if there is a shortfall. If any system errors occur during the translation, all pre-deducted general translation quotas will be fully refunded to your account. When the automatic term extraction feature is not enabled, the pre-deduction per page is 3000 Tokens/page. When the automatic term extraction feature is enabled, the pre-deduction per page is 6000 Tokens/page.<br/><br/>From actual usage statistics, when automatic term extraction is enabled, the average consumption is about 7381.56 Tokens per page, with the P25–P75 range being 4215.42–9490.89 Tokens; when automatic term extraction is disabled, the average consumption is about 4702.98 Tokens per page, with the P25–P75 range being 2642.38–5797.40 Tokens. Here, P25 (the 25th percentile) indicates that 25% of the pages consume fewer Tokens than this value, and P75 (the 75th percentile) indicates that 75% of the pages consume fewer Tokens than this value. The P25–P75 range is used to describe the common consumption range for most pages, thereby reducing the impact of extreme samples on the overall assessment.
我們目前採用基於「翻譯額度」的計費方式，具體說明如下：

<ul><li>翻譯額度按照翻譯過程中使用的 <b> Token 數量 </b> 進行計費。</li><li><b> GLM-4-Flash-250414 </b> 模型所有用戶可用，所有 BabelDOC 用戶每月可獲得 50 萬 Token 額度，用來在 BabelDOC 中使用免費模型翻譯文件。（實際可用性取決於上游供應商的供應情況，額度根據伺服器負載動態調整）</li><li><b> Qwen3.5-Plus/DeepSeek-V3.24-Flash/doubao-seed-2-0-lite/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7 </b> 模型僅限 <b> Pro/Max </b> 會員使用，計入會員通用翻譯額度，Pro 會員每月可享受 <b> 高達 2000 萬 Token </b> 的翻譯額度，Max 會員每月可享受 <b> 高達 5000 萬 Token </b> 的翻譯額度。失敗的翻譯不計入翻譯額度。</li></ul>

在翻譯任務啟動時，系統將根據"3000 Token × 翻譯頁數"的計算方式預先扣除相應的通用翻譯額度。待翻譯完成後，我們將按實際消耗進行結算，多餘額度將退還至您的賬戶，不足部分則需補充。若翻譯過程中遇到任何系統錯誤，所有預扣的通用翻譯額度將自動全額返還至您的賬戶。當不啟用自動提取術語時，單頁預扣 Token 數為 3000 Token/頁。當啟用自動提取術語功能時，單頁預扣 Token 數為 6000 Token/頁。<br/><br/>從實際使用統計來看，啟用自動術語提取時，平均每頁約消耗 7381.56 Token，其中 P25–P75 區間為 4215.42–9490.89 Token；關閉自動術語提取時，平均每頁約消耗 4702.98 Token，P25–P75 區間為 2642.38–5797.40 Token。其中，P25（第 25 百分位）表示有 25% 的頁面 Token 消耗低於該值，P75（第 75 百分位）表示有 75% 的頁面 Token 消耗低於該值，P25–P75 區間可用於描述大多數頁面的常見消耗範圍，從而減少極端樣本對整體評估的干擾。
13 days ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
We currently use a "translation quota" based billing method, detailed as follows:<br/><br/><ul><li>Translation quotas are charged based on the number of <b>Tokens</b> used during the translation process.</li><li>The <b>GLM-4-Flash-250414</b> model is available to all users, and every BabelDOC user receives a monthly quota of 500,000 Tokens for free model translations within BabelDOC. (Actual availability depends on the supply from upstream suppliers, and quotas are dynamically adjusted based on server load)</li><li>Models such as <b>Qwen3.5-Plus/DeepSeek-V4-Flash/doubao-seed-2-0-lite/GPT-5-mini/Gemini-3-Flash/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7</b> are exclusively for <b>Pro/Max</b> members, counted towards the members' general translation quota. Pro members enjoy up to <b>20 million Tokens</b> per month, while Max members can enjoy up to <b>50 million Tokens</b> per month. Failed translations are not counted towards the translation quota.</li></ul><br/>At the start of a translation task, the system will deduct the general translation quota in advance based on the calculation of "3000 Tokens × number of pages". Once the translation is complete, we will settle based on actual consumption, refunding any excess to your account, or requiring additional if insufficient. If any system errors occur during the translation, all deducted general translation quotas will be fully refunded to your account. When the automatic term extraction feature is not enabled, the deduction per page is 3000 Tokens/page. When the automatic term extraction feature is enabled, the deduction per page is 6000 Tokens/page.<br/><br/>From actual usage statistics, when automatic term extraction is enabled, the average consumption is about 7381.56 Tokens per page, with the P25–P75 range being 4215.42–9490.89 Tokens; when automatic term extraction is disabled, the average consumption is about 4702.98 Tokens per page, with the P25–P75 range being 2642.38–5797.40 Tokens. Here, P25 (the 25th percentile) indicates that 25% of the pages consume fewer Tokens than this value, and P75 (the 75th percentile) indicates that 75% of the pages consume fewer Tokens than this value. The P25–P75 range is used to describe the common consumption range for most pages, thereby reducing the impact of extreme samples on the overall assessment.
我們目前採用基於「翻譯額度」的計費方式，具體說明如下：

<ul><li>翻譯額度按照翻譯過程中使用的 <b> Token 數量 </b> 進行計費。</li><li><b> GLM-4-Flash-250414 </b> 模型所有用戶可用，所有 BabelDOC 用戶每月可獲得 50 萬 Token 額度，用來在 BabelDOC 中使用免費模型翻譯文件。（實際可用性取決於上游供應商的供應情況，額度根據伺服器負載動態調整）</li><li><b> Qwen3.5-Plus/DeepSeek-V3.24-Flash/doubao-seed-2-0-lite/GPT-5-mini/Gemini-3-Flash/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7 </b> 模型僅限 <b> Pro/Max </b> 會員使用，計入會員通用翻譯額度，Pro 會員每月可享受 <b> 高達 2000 萬 Token </b> 的翻譯額度，Max 會員每月可享受 <b> 高達 5000 萬 Token </b> 的翻譯額度。失敗的翻譯不計入翻譯額度。</li></ul>

在翻譯任務啟動時，系統將根據"3000 Token × 翻譯頁數"的計算方式預先扣除相應的通用翻譯額度。待翻譯完成後，我們將按實際消耗進行結算，多餘額度將退還至您的賬戶，不足部分則需補充。若翻譯過程中遇到任何系統錯誤，所有預扣的通用翻譯額度將自動全額返還至您的賬戶。當不啟用自動提取術語時，單頁預扣 Token 數為 3000 Token/頁。當啟用自動提取術語功能時，單頁預扣 Token 數為 6000 Token/頁。<br/><br/>從實際使用統計來看，啟用自動術語提取時，平均每頁約消耗 7381.56 Token，其中 P25–P75 區間為 4215.42–9490.89 Token；關閉自動術語提取時，平均每頁約消耗 4702.98 Token，P25–P75 區間為 2642.38–5797.40 Token。其中，P25（第 25 百分位）表示有 25% 的頁面 Token 消耗低於該值，P75（第 75 百分位）表示有 75% 的頁面 Token 消耗低於該值，P25–P75 區間可用於描述大多數頁面的常見消耗範圍，從而減少極端樣本對整體評估的干擾。
13 days ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
All users currently have access to the GLM-4-Flash-250414 model for translation, while Pro users can also choose to use the Qwen3.5-Plus/DeepSeek-V4-Flash/doubao-seed-2-0-lite/GPT-5-mini/Gemini-3-Flash/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7 models for translation, subject to actual availability.

After comparison, we found that Kimi-K2.5 is effective in extracting terminology, hence we are introducing the "Kimi + Qwen" and "Kimi + DeepSeek" translation services:

If you choose the "Kimi + Qwen" translation service and enable automatic terminology extraction, Kimi-K2.5 will be used to extract terminology, and Qwen3.5-Plus will be used to complete the text translation. All related consumption will be counted towards Qwen usage. (Limited time test)

If you choose the "Kimi + DeepSeek" translation service and enable automatic terminology extraction, Kimi-K2.5 will be used to extract terminology, and DeepSeek-V4-Flash will be used to complete the text translation. All related consumption will be counted towards DeepSeek usage. (Limited time test)

Since BabelDOC uses a backend translation architecture, all translation logic is executed on the server side, independent of the translation engine used by the plugin. Therefore, the range of translation engines available for BabelDOC differs from those available for immersive translation plugins, and BabelDOC only supports certain advanced translation models, not all advanced models are available.
所有用戶目前均可使用 GLM-4-Flash-250414 模型進行翻譯，而 Pro 用戶還可額外選擇使用 Qwen3.5-Plus/DeepSeek-V3.24-Flash/doubao-seed-2-0-lite/GPT-5-mini/Gemini-3-Flash/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7 模型（以實際可選模型為準）進行翻譯。

經過對比我們發現 Kimi-K2.5 提取術語效果較好，因此推出“Kimi + Qwen”和“Kimi + DeepSeek”翻譯服務：

若選擇“Kimi + Qwen”翻譯服務，且啟用自動術語提取，則會使用 Kimi-K2.5 提取術語，並使用 Qwen3.5-Plus 完成文本翻譯。相關消耗均會計入 Qwen 消耗。（限時測試）

若選擇“Kimi + DeepSeek”翻譯服務，且啟用自動術語提取，則會使用 Kimi-K2.5 提取術語，並使用 DeepSeek-V3.24-Flash 完成文本翻譯。相關消耗均會計入 DeepSeek 消耗。（限時測試）

由於 BabelDOC 採用後端翻譯架構，所有翻譯邏輯均在伺服器端執行，與插件所使用的翻譯引擎相互獨立。因此，BabelDOC 可選用的翻譯引擎範圍與沉浸式翻譯插件存在差異，BabelDOC 僅支持部分高級翻譯模型，並非所有高級模型均可使用。
13 days ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
All users can currently use the GLM-4-Flash-250414 model for translation free of charge, while Pro users can also choose to use the Qwen3.5-Plus/DeepSeek-V4-Flash/doubao-seed-2-0-lite models for translation.
所有用戶目前均可使用 GLM-4-Flash-250414 模型進行翻譯，而 Pro 用戶還可額外選擇使用 Qwen3.5-Plus/DeepSeek-V3.24-Flash/doubao-seed-2-0-lite/Kimi+Qwen/Kimi+DeepSeek/GLM-4.7 模型（以實際可選模型為準）進行翻譯。

經過對比我們發現 Kimi-K2.5 提取術語效果較好，因此推出“Kimi + Qwen”和“Kimi + DeepSeek”翻譯服務：

若選擇“Kimi + Qwen”翻譯服務，且啟用自動術語提取，則會使用 Kimi-K2.5 提取術語，並使用 Qwen3.5-Plus 完成文本翻譯。相關消耗均會計入 Qwen 消耗。（限時測試）

若選擇“Kimi + DeepSeek”翻譯服務，且啟用自動術語提取，則會使用 Kimi-K2.5 提取術語，並使用 DeepSeek-V3.24-Flash 完成文本翻譯。相關消耗均會計入 DeepSeek 消耗。（限時測試）

由於 BabelDOC 採用後端翻譯架構，所有翻譯邏輯均在伺服器端執行，與插件所使用的翻譯引擎相互獨立。因此，BabelDOC 可選用的翻譯引擎範圍與沉浸式翻譯插件存在差異，BabelDOC 僅支持部分高級翻譯模型，並非所有高級模型均可使用。
13 days ago
User avatar None

Resource updated

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
The “app/src/i18n/zh-HK.json” file was changed. 2 weeks ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
The current version has the following limitations, which we are actively optimizing:
<ol>
<li>Some formula sections may overlap.</li>
<li>Italicized text in translations will disappear.</li>
<li>Small sections (such as author information, citations, table of contents entries, etc.) may not be accurately recognized.</li>
<li>Only supports languages written in Latin, Chinese, Japanese, or Korean scripts as the source language.</li>
<li>Does not support documents with Vietnamese as the source language.</li>
<li>Does not support underlining.</li>
<li>Does not support translation of chart areas.</li>
<li>Does not support drop caps.</li>
<li>Does not support PDF Outline translation.</li>
<li>Does not support scanned, image-based, or OCR-based PDFs.</li>
<li>Does not support ligatures.</li>
<li>Does not support blueprints.</li>
<li>Does not support Type3 fonts.</li>
<li>Does not support image format inline formulas.</li>
<li>Does not support right-to-left formatted documents.</li>
<li>Only supports horizontally aligned text.</li>
<li>Currently, selective ignoring of translation areas is not supported (the system will default to ignoring the translation of reference areas)</li></ol>
當前版本存在以下限制，我們正在積極優化：<ol><li>部分公式段落可能出現重疊</li><li>譯文中傾斜的文本會消失</li><li>小型段落（如作者信息、引文、目錄條目等）可能識別不準</li><li>僅支持使用拉丁字母或中日韓文書寫的語言作為原語言</li><li>不支持越南語作為原語言的文件</li><li>不支持下劃線</li><li>不支持圖表區域翻譯</li><li>不支持首字下沉效果</li><li>不支持 PDF Outline 翻譯</li><li>不支持掃描版、圖片版及 OCR 版 PDF</li><li>不支持連字</li><li>不支持圖紙</li><li>不支持使用 Type3 字體</li><li>不支持圖片格式的行內公式</li><li>不支持從右到左排版的文件</li><li>僅支持水平橫排文本</li></ol>
2 weeks ago
User avatar None

String added in the repository

  1. Immersive Translate Site
  2. Immersive Translate Document App
  3. Chinese (Traditional, Hong Kong)
Please enable your installed Immersive Translate extension to continue.
請啟用已安裝的沉浸式翻譯擴充功能，以便正常使用
2 weeks ago
User avatar None

Resource updated

  1. Immersive Translate Site
  2. Global i18n
  3. Chinese (Traditional, Hong Kong)
The “global-i18n/en-US.json” file was changed. a month ago
User avatar None

String updated in the repository

  1. Immersive Translate Site
  2. Global i18n
  3. Chinese (Traditional, Hong Kong)
Month
月
a month ago
Browse all changes for this language