Translation tool: Translate 100+ Languages Automatically

The Translation tool uses advanced language identification algorithms trained on billions of multilingual text samples. Simply omit the source language parameter, and the tool analyzes linguistic patterns, character sets, and statistical markers to identify the input language. Detection accuracy reaches 97% for texts exceeding 50 characters and improves to 99.5% for texts over 200 characters. The system recognizes all major languages plus regional variants (e.g., Brazilian Portuguese vs. European Portuguese, Simplified Chinese vs. Traditional Chinese).

Every auto-detected language includes a confidence score from 0 to 1.0. High-confidence detections (>0.9) proceed automatically, while lower scores trigger fallback mechanisms. Mixed-language texts return multiple language candidates with individual confidence scores, allowing you to choose the most appropriate source language. The API also detects code-switching (language mixing within a single text) and can segment translations accordingly.

When auto-detection confidence falls below your specified threshold, the tool can either reject the request with a detailed error, fall back to a default language you specify, or attempt translation using the highest-confidence candidate with a warning flag. This flexibility ensures you never lose data while maintaining translation quality standards.

Batch translation distributes workload across multiple translation engines simultaneously, reducing total processing time by up to 90% compared to sequential translation. Submit an array of up to 1000 texts in a single API call, and the system automatically partitions the workload, processes segments in parallel, and reassembles results in the original order. Performance scales linearly: 100 texts complete in ~5 seconds, 500 texts in ~20 seconds, and 1000 texts in ~45 seconds.

Within a batch request, the Translation tool maintains a translation memory cache to ensure terminological consistency. When the same phrase appears multiple times across different texts, the tool reuses the same translation, preserving consistency. This is particularly valuable for product catalogs, UI strings, and documentation where identical terms must translate uniformly. Translation memory operates automatically without configuration and persists only for the duration of the single request to protect data privacy.

Upload custom glossaries to enforce specific translations for brand names, technical terms, or industry jargon. Glossaries support up to 10,000 term pairs per language pair and apply across all translations in a request. Define terms once, then reference the glossary ID in subsequent API calls. The system prioritizes glossary entries over neural translation suggestions, guaranteeing 100% consistency for specified terms. Glossaries are particularly effective for maintaining brand voice, translating acronyms, and preserving product names.

Control the tone of translations with three formality settings: informal (casual conversations, social media), formal (business communications, official documents), and technical (API documentation, scientific papers). Formality affects pronoun usage (tu vs. vous in French), verb conjugations, and vocabulary selection. The default 'auto' mode analyzes source text tone and matches it in the target language, ensuring appropriate register for each context.

Specify the subject domain to activate specialized translation models trained on domain-specific corpora. Available domains include legal (contracts, regulations), medical (clinical notes, pharmaceutical), marketing (ads, brand messaging), technical (engineering, software), financial (reports, statements), and general (default). Domain models recognize field-specific terminology and apply context-appropriate translations. For example, the medical domain correctly translates 'cold' as a symptom rather than temperature.

Long texts are segmented into manageable chunks for translation, but this can lose inter-sentence context. Enable context window preservation to provide surrounding sentences as context when translating each segment. The API maintains a sliding window of up to 3 preceding sentences, helping the model resolve ambiguous pronouns, maintain narrative consistency, and select contextually appropriate word senses. This feature is essential for translating articles, stories, and long-form content.

blogTranslationApi.contextPreservation.terminology.content

blogTranslationApi.contextPreservation.formatting.content

blogTranslationApi.contextPreservation.placeholders.content

Translate a single text with automatic language detection

blogTranslationApi.implementation.basicUsage.code

Translate multiple texts efficiently in a single request

blogTranslationApi.implementation.batchProcessing.code

Implement robust error handling for production reliability

blogTranslationApi.implementation.errorHandling.code

Cache results and reuse glossaries for faster processing

blogTranslationApi.implementation.optimization.code

• Translate 5000 product descriptions (average 150 words each)

• Target languages: German, French, Spanish, Italian, Portuguese, Dutch, Swedish, Polish, Czech, Romanian

• blogTranslationApi.realWorldExample.requirements.req3

• Maintain consistency for 200 product-specific terms (materials, features, brand names)

• Maintain informal, marketing-friendly tone across all languages

• Complete translation within 24 hours to meet launch deadline

• Created a custom glossary with 200 product terms mapped to each target language

• Extracted product descriptions from database with HTML intact

• Batched translations into groups of 500 descriptions per API call

• Used parallel processing: 10 API calls (one per language) running simultaneously

• Applied 'marketing' domain and 'informal' formality to match brand voice

• Validated HTML preservation with automated testing on sample outputs

blogTranslationApi.realWorldExample.implementation.code