Description
Amazon Translate is AWS’s neural machine translation service, delivering real-time and batch translation across 75-plus languages through an API. It supports real-time translation of text, HTML, and DOCX documents while preserving the original formatting, plus asynchronous batch translation for large document sets stored in Amazon S3. Custom Terminology lets a business define how brand names and industry-specific terms get translated, up to 10,000 terms per file at no extra cost, and Active Custom Translation improves output by learning from provided translation examples without training a dedicated model. Amazon Translate integrates directly with other AWS services like Lambda, S3, and Comprehend for end-to-end translation workflows.
Key Features
- 75+ languages — real-time and batch neural machine translation across thousands of language pairs.
- Real-time document translation — translates Text, HTML, and DOCX files while preserving the original format.
- Batch document translation — asynchronous translation of large document sets stored in Amazon S3.
- Custom Terminology — defines translation of brand names and industry terms, up to 10,000 terms per file at no extra cost.
- Active Custom Translation — improves output using provided translation examples without training a custom model.
- Automatic language detection — identifies the source language for content like reviews or social posts with no language code.
How It Works
A developer sends text or a document to Amazon Translate through the AWS Console, CLI, or SDK, and it returns a neural machine translation in the target language, preserving formatting for real-time document translation. Large volumes of pre-existing documents can instead be submitted as an asynchronous batch job, translating an entire folder of files stored in S3 through a single API call. Custom Terminology and Active Custom Translation let a business steer output toward its own brand language and preferred terminology without training and maintaining a dedicated translation model. Everything runs on a pay-per-character basis with no contracts or minimum commitments, and integrates directly with other AWS services like Lambda, S3, and Comprehend.
Technical Architecture & Overview
- Core Engine: AWS’s proprietary neural machine translation models.
- Deployment: AWS Console, CLI, and SDK access; a browser plugin is available for on-page translation.
- API Surface: REST API billed per character, with SSL encryption in transit and encryption at rest.
- Known Limits: The 2 million character/month free tier expires after 12 months; custom model training and hosting are billed separately from standard translation.
Pros & Cons
| Pros | Cons |
|---|---|
| Pay-per-character pricing with no contracts or minimum commitments keeps costs proportional to actual usage. | The free tier expires after 12 months rather than continuing indefinitely. |
| Custom Terminology supports up to 10,000 terms per file at no extra cost. | Custom model training and hosting are billed separately from standard translation usage. |
| Real-time document translation preserves the original file’s formatting. | Getting started requires familiarity with the broader AWS ecosystem and IAM permissions. |
| Deep integration with AWS services like Lambda, S3, and Comprehend simplifies workflows already built on AWS. | There’s no consumer-facing app; Amazon Translate is API and console-only for developers. |
| The free tier gives 2 million characters a month for 12 months, generous for early testing. | Pricing can become significant at high volume compared to some flat-rate competitors. |
Pricing
| Plan | Price |
|---|---|
| Free Tier | $0 for first 2M characters/month (12 months) |
| Standard Translation | $15 per million characters |
| Active Custom Translation | $15 per million characters + example data costs |
Platform Availability
AWS Console | CLI | SDK | API | Browser Plugin
Best For
Developers building multilingual apps already on AWS infrastructure | Businesses needing batch translation of large document sets | Teams needing brand-consistent terminology across translations
Similar Tools in This Directory
Also worth comparing: Azure AI Translator, Microsoft’s pay-per-character translation API | Google Translation AI, Google Cloud’s neural translation API and AutoML custom models | DeepL Translator, a translation tool known for natural-sounding European language output | Phrase Language AI, an enterprise localization platform with AI-driven engine selection
Visit Official Website
Frequently Asked Questions
Does Amazon Translate have a free tier?
Yes. New AWS accounts get 2 million characters of standard translation free per month for the first 12 months, according to AWS’s official pricing page.
How much does Amazon Translate cost?
Standard Text Translation costs $15 per million characters after the free tier, according to AWS’s official pricing page.
Does Amazon Translate support document translation?
Yes. It translates Text, HTML, and DOCX documents in real time while preserving the original formatting, and supports batch translation of larger file sets via Amazon S3, according to AWS’s official site.
Can Amazon Translate learn custom terminology?
Yes. Custom Terminology lets a business define how up to 10,000 brand or industry-specific terms per file are translated, at no additional cost, according to AWS’s official site.
Does Amazon Translate detect the source language automatically?
Yes. It can automatically identify the source language with high accuracy for content that has no language code, according to AWS’s official site.
