Amazon Translate
Machine translation on AWS: real-time text and document translation, asynchronous batch jobs from S3, automatic source language detection, custom terminology, parallel data with Active Custom Translation, formality and profanity settings, security, cost, and pipelines with Transcribe and Comprehend.
3 min read · Last reviewed · SAA-C03
Sign in to bookmark tutorials and track what you have read.
On this page
What it is
Amazon Translate is a machine translation service that translates text from one language to another on demand, using neural machine translation. Use it to translate unstructured text and documents, or to build applications that work in many languages.
Choose Translate when a question asks to translate content, chat, emails, product listings, subtitles, or support tickets, so that customers and employees can use their preferred language.
Features
- Real-time translation of text and documents, with immediate responses through the console or API (Application programming interface)A defined way for one program to ask another for data or actions, such as a set of HTTPS requests and responses. Every AWS service is controlled through its API, which the console, the CLI, and the SDKs call for you.Glossary entry.
- Asynchronous batch translation: translate large collections of documents (up to 5 GB) stored in Amazon S3 with
StartTextTranslationJob, writing the results back to S3. This suits collections of short documents, such as social media posts or user reviews, when immediate results aren't needed. - Automatic source language detection: set the source language to
autoand Translate identifies it (this may add a charge). - Custom terminology: make sure brand names, product names, and technical terms are translated the way you want, or left untranslated.
- Parallel data (Active Custom Translation): provide example translations so batch jobs match your style and domain, without building a custom model.
- Formality setting: choose formal or informal language for supported target languages.
- Profanity masking: replace profane words in the output with a grawlix string ("?$#@$").
Combining with other services
- Transcribe → Translate → Polly: speech in one language, read aloud in another.
- Translate → Comprehend: translate social media or reviews, then analyze sentiment and entities.
- Transcribe → Translate: multilingual subtitles and live captions.
Security, resilience, performance, and cost
Security
- IAM controls access; batch jobs use an IAM role to read from and write to S3.
- Encrypt batch output with a KMS key; traffic uses TLS (Transport Layer Security)The protocol that encrypts traffic between a client and a server and lets the server prove its identity with a certificate. It replaced SSL, though the older name is still common; HTTPS is HTTP over TLS.Glossary entry.
- CloudTrail records API calls.
Resilience and performance
- Translate is a managed, Regional service. Use real-time calls for interactive apps and batch jobs for large volumes.
Cost
- Pay per character translated. Active Custom Translation with parallel data is priced higher than standard translation.
Monitoring
- CloudWatch metrics for requests, characters, and errors; CloudTrail for auditing.
Translate or another option?
| If the requirement is | Choose |
|---|---|
| Translate text or documents between languages | Amazon Translate |
| Keep brand and product names consistent | Translate custom terminology |
| Match our own style in batch translations | Parallel data (Active Custom Translation) |
| Detect the language of text, without translating | Amazon Comprehend |
| Convert speech to text first | Amazon Transcribe |
| Speak the translated text | Amazon Polly |
Reverse mindmap: when the question points here
- Yes, when the question says
- Translate text, chat, or documents leads to Translate
- Multilingual website or app content leads to Translate
- Translate a large set of documents in S3 leads to Batch translation job
- Tune the output
- Brand names translated correctly leads to Custom terminology
- Match our style and domain leads to Parallel data
- Formal or informal tone leads to Formality setting
- Hide profanity leads to Profanity masking
- Combine it
- Speech in one language, audio in another leads to Transcribe, Translate, Polly
- Sentiment of foreign-language reviews leads to Translate, then Comprehend
- No, when the question says
- Only detect the language leads to Comprehend
Key takeaways
- Translate provides neural machine translation for text and documents, in real time or as batch jobs from S3 (up to 5 GB).
- It can detect the source language automatically.
- Custom terminology fixes specific terms; parallel data adapts batch output to your style; formality and profanity settings refine the output.
- It is priced per character and often chained with Transcribe, Comprehend, and Polly.
Sources
- Amazon Translate Developer Guide: What is Amazon Translate?
- Amazon Translate Developer Guide: Real-time translation
- Amazon Translate Developer Guide: Asynchronous batch processing
- Amazon Translate Developer Guide: Supported languages and language codes
- Amazon Translate Developer Guide: Customizing your translations
- Amazon Translate Developer Guide: Customizing your translations with custom terminology
- Amazon Translate Developer Guide: Customizing your translations with parallel data
- Amazon Translate Developer Guide: Setting formality
- Amazon Translate Developer Guide: Masking profane words and phrases
- Amazon Translate Developer Guide: Encryption at rest
- Amazon Translate Developer Guide: Logging Amazon Translate API calls with AWS CloudTrail
- Amazon Translate pricing