Translation
Text translation
Translate text from the dashboard with language, tone, glossary, and Translation Memory controls. Input accepts up to 100,000 characters, subject to the model token budget.
Paste text into the dashboard translator and choose the target language. Input accepts up to 100,000 characters per request, subject to the model token budget. The text translator uses the same tone, glossary, and Translation Memory settings as a document job.
How it works
Step by step with screenshots: Text translation, tones and glossary.
- Auto-detect recognises the source language from the text itself; 42 target languages are supported.
- Tone (Standard plus nine alternatives) and glossary are optional, and the choices are remembered for next time.
- The translation appears progressively as the model generates it. Line breaks are preserved; rich document formatting belongs in the document workflow.
- Changing the tone requests a new translation of the same source.
Translation memory reuse
If translation memory is enabled and you've translated the exact same text before, we return your stored translation instead of generating a new one. It's faster, it's consistent with your previous work, and exact TM matches don't generate new model output. See Translation memory for how storage works.
The same settings apply to short text
A short request can still contain an approved product name or a sentence translated before. Keeping the document settings available here means you do not need a separate process for a paragraph copied from an email, ticket, or draft.
Save your language pair, tone, and glossary as a preset. Tone controls translation style. A domain saved in a preset is used only when extracting terminology.
Keep exploring
Tone
Tone controls the register of the result: nine tones on top of the standard voice.
Glossary
Store approved term pairs for a language pair and apply them during text translation, document translation, and review.
Translation memory
Save approved source and target segments, reuse exact matches, and review fuzzy matches when similar text appears.