Gcp Translation

Google Translate API (V3) provides high-accuracy neural machine translation across 100+ languages. For agents: translate user input to English for LLM processing then translate output back, detect source language automatically, translate documents (PDF, DOCX) while preserving formatting, or use glossaries to handle domain-specific terminology consistently. Confidence is docs-derived.

Python SDK V3: from google.cloud import translate_v3 as translate. client = translate.TranslationServiceClient(). parent = f"projects/{project}/locations/global". response = client.translate_text(parent=parent, contents=["Hello, world!"], mime_type="text/plain", source_language_code="en-US", target_language_code="fr"). response.translations[0].translated_text. Language detection: client.detect_language(parent=parent, content="Bonjour!"). Document translation: client.translate_document(parent=parent, document_input_config={...}, target_language_code="es").

Service account or ADC. Roles: roles/cloudtranslate.user (translate operations). HTTPS enforced. Request content processed by Google servers.

InvalidArgumentError for unsupported language codes. RESOURCE_EXHAUSTED for quota limits. Translation API SLA: 99.9%. Accuracy is best-in-class for major language pairs; lower for less-resourced languages. Document translation: preserves formatting for PDF/DOCX. Batch translation available for large volumes via GCS.

cloud.google.com/translate/docs covers V2 and V3 APIs, document translation, glossaries, and batch translation. Getting started: enable Translation API, ADC configured, first translate in under 5 minutes. Free tier: 500K characters/month (V2); V3 billed per character. SDK for Python, Java, Go, Node.js. Community via Google Cloud Community.