Amazon Translate
Neural machine translation across 75+ languages.
❓ What is it?
A neural machine-translation service for real-time and batch translation, with custom terminology to lock brand/domain terms, formality control in supported languages, and profanity masking.
💡 Why does it exist?
Serving global users in their language multiplies reach, and human translation cannot keep pace with dynamic content like chats, reviews, and tickets. Machine translation makes multilingual the default rather than a project.
⏱️ When should you use it?
Use it to localise user-generated content, translate support conversations live, or batch-translate document sets; add custom terminology whenever product names must survive translation untouched.
🗺️ Where does it fit?
Inline in text pipelines: between Transcribe and Polly for speech-to-speech translation, in front of Comprehend for analysing foreign-language text, or inside chat backends for live localisation.
🔌 How do you integrate it?
Call TranslateText for synchronous needs or StartTextTranslationJob for S3 batches; attach a terminology file (CSV/TMX) to control key phrases.
🧩 Commonly integrated with
🎯 Exam angle (AIF-C01)
- The speech-to-speech chain to memorise: Transcribe → Translate → Polly.
- Custom terminology answers "brand names must not be translated".