DeepL API
The DeepL API lets developers easily add high-quality AI translation to their own projects.
DeepL consistently outperforms other machine translation providers in blind quality tests, particularly for European languages. The API supports 33 languages including German, French, Spanish, Portuguese, Japanese, Chinese, and more. Features include document translation (Word, PDF, HTML), custom glossaries for technical terminology, and formality level control. Pricing starts at $25/month for the Pro plan (1 million characters). The free tier at 500K chars/month is generous. First choice for European enterprises and publishers requiring the highest translation quality.
API Details
Categories
Frequently Asked Questions
DeepL consistently outperforms Google Translate in blind tests for European language pairs, particularly for nuanced, formal, and technical texts. DeepL produces more natural-sounding translations with better sentence structure. However, Google Translate supports more languages (130+ vs 33 for DeepL). For European languages where quality matters, DeepL is generally the professional choice.
DeepL API Free allows 500,000 characters per month at no cost. DeepL API Pro starts at u20ac5.99/month for 500,000 characters, then u20ac20 per million additional characters. Pro includes higher rate limits, priority processing, and CAT tool integrations. For high-volume use, DeepL Business offers volume discounts.
DeepL currently supports 33 languages including all major European languages (English, German, French, Spanish, Italian, Dutch, Polish, Portuguese, Russian) plus Chinese, Japanese, Korean, Indonesian, and Arabic. DeepL quality is particularly strong for German-English and French-English. New languages are added regularly.
Yes. The DeepL API Pro tier supports document translation for DOCX, PPTX, PDF, XLSX, HTML, and TXT files while preserving the original formatting. This is extremely useful for translating business documents. Document translation is charged at the same per-character rate as text. Note that PDFs must be text-based (not scanned images).
