Document translation

Updated at:

This topic introduces Document Translation, a feature of Multimodal Translation within Alibaba Cloud Machine Translation.

Note

For more product-related questions, visit pre-sales consultation or search for the DingTalk group ID 23369411 to consult with experts.

Product overview

You can integrate Document Translation by using an API or access it through an online editing platform. It supports a wide range of document formats and preserves the original layout with high fidelity after translation. You can choose from machine translation engines optimized for various fields, including general, e-commerce, social, medical, and technology, to ensure accurate, stable, and efficient document translation services.

Use cases

  • Medical literature

  • Academic papers

  • Cross-border transaction contracts

  • Immigration and customs documents

Advantages

  • Easy-to-use API

Our standardized API significantly reduces your development effort.

  • Format preservation

The service preserves the original document format with high fidelity, greatly improving efficiency and user experience.

  • Advanced algorithms

It is powered by a state-of-the-art, end-to-end deep learning framework trained on hundreds of millions of parallel sentence pairs.

  • Multi-domain support

Our translation engines are optimized for multiple domains, including general, e-commerce, and medical, to meet the translation needs of specialized documents.

Supported formats

1. Online editing platform

doc, docx, xls, xlsx, csv, ppt, pptx, txt, srt, html, xml, xhtml, json, xliff, tmx, dita, md, markdown, idml, pdf

2. API

doc, docx, xls, xlsx, csv, ppt, pptx, txt, srt, html, htm, xml, xhtml, json, xliff, tmx, dita, md, markdown, idml, pdf

Document size limits

1. Online editing platform

The maximum file size is 50 MB. The content limit is 1 million characters for doc, docx, csv, txt, srt, pdf, html, htm, xml, xhtml, json, xliff, tmx, dita, md, markdown, and idml files. For xls and xlsx files, the limit is 100,000 cells.

2. API

The maximum file size is 50 MB. The content limit is 1 million characters for doc, docx, csv, txt, srt, pdf, html, htm, xml, xhtml, json, xliff, tmx, dita, md, markdown, and idml files. For xls and xlsx files, the limit is 100,000 cells.

Supported scenarios

The service supports general, e-commerce, medical, social, and technology scenarios, leveraging tens of millions of bilingual corpora and millions of professional bilingual terms.

1. Online editing platform

Scenario

Scenario module

Scenario code

General

--

general

Technology

--

tech

2. API

Scenario

Scenario module

Scenario code

General

--

general

E-commerce

Product title

title

Product description

description

Communication between buyers and sellers

message

Medical

--

medical

Social

--

social

Technology

--

tech

Supported languages

1. Online editing platform

  • General scenario

Bidirectional translation is supported between any two of the following 12 languages: Chinese, English, Korean, Japanese, French, Spanish, Portuguese, Italian, Russian, Arabic, Indonesian, and Vietnamese.

The following three language pairs are also supported:

Source language

Target language

Chinese

Traditional Chinese

Traditional Chinese

Chinese

Chinese

Cantonese

  • In the technology scenario, the following two language pairs are supported:

Source language

Target language

Chinese

English

English

Japanese

2. API

  • General scenario, with 243 language pairs supported.

Bidirectional translation is supported between any two of the following 16 languages: Chinese, English, Korean, Japanese, French, Spanish, Portuguese, Italian, Russian, Arabic, Turkish, Indonesian, Vietnamese, Thai, German, and Polish.

The following three language pairs are also supported:

Source language

Target language

Chinese

Traditional Chinese

Traditional Chinese

Chinese

Chinese

Cantonese

  • In the technology scenario, the following two language pairs are supported:

Source language

Target language

Chinese

English

English

Japanese

  • In the medical scenario, the following four language pairs are supported:

Source language

Target language

Chinese

English

English

Chinese

Chinese

Japanese

Japanese

Chinese

  • In the social scenario, the following six language pairs are supported:

Source language

Target language

Chinese

English

English

Chinese

Chinese

Japanese

Japanese

Chinese

Chinese

Korean

Korean

Chinese

  • In the e-commerce scenario, the following 35 language pairs are supported:

Source language

Target language

Chinese

English

English

Chinese

Chinese

Japanese

Japanese

Chinese

Chinese

Korean

Korean

Chinese

Chinese

Italian

Chinese

Thai

Chinese

Vietnamese

Chinese

Malay

Malay

Chinese

Chinese

Traditional Chinese

Chinese

Cantonese

Chinese

Spanish

Chinese

French

Chinese

Russian

English

Russian

English

Japanese

English

Arabic

English

Hindi

English

Hebrew

English

Polish

English

Dutch

English

Turkish

English

French

English

Spanish

English

Italian

English

German

English

Thai

English

Indonesian

English

Vietnamese

English

Malay

Malay

English

Japanese

English

English

Korean

3. Language code list

When calling the API, you must provide the language codes for the source language and target language. Ensure that you use the correct language codes to avoid call failures. For more information, see Language Codes.

Online document translation and editing platform

Document Translation provides an online document translation and editing platform that allows non-developers to translate and edit documents online. After you activate the Document Translation service, go to the Machine Translation console, find the Document Translation product card, and click Online Editing Platform to access the platform and its features.

For more information about using the platform, see User guide for the online document translation and editing platform.

API quick reference

For more information about API calls, see the Document Translation Calling Guide.

Note

For more usage-related questions, visit pre-sales consultation or search for the DingTalk group ID 23369411 to consult with experts.