How it works
Three steps. One output you can actually use.
Detect
OCR scans every text region in the image, with bounding box coordinates and confidence scores per word.
Remove & Translate
The original text is cleanly removed from the image. Then each region is translated into your target language.
Render back
Translated text is rendered onto the image in the exact same position. Download a finished image — not a transcript.
Three outputs per image
Every translation returns all three — use whichever you need.
Translated image
The final output. Original image with translated text rendered back in position. Ready to publish, share, or print.
Text-removed image
The image with all text erased and background reconstructed. Use as a clean base if you want to apply your own typography.
Region breakdown
Every detected text region with original text, translation, position, and OCR confidence score — for validation or downstream automation.
See it in action
Upload an image with text — get a translated image back, with text rendered in the same position.
Input
Image with German text
Output
Same image, text translated to English
Text is detected, removed, translated, and rendered back — in exactly the same position. You download a finished image.
Built for scale
Batch processing
Upload up to 20 images and translate to up to 10 languages in one job. Real-time progress, cancel any time.
Full API access
Generate API keys and integrate image translation directly into your pipeline. Webhooks supported for async batch jobs.
11 European languages
English, German, French, Spanish, Italian, Portuguese, Dutch, Swedish, Danish, Norwegian, Finnish.
Exclude specific text
Keep logos, brand names, handles, or any pattern untranslated. Your brand stays intact.
Confidence scores
Every OCR region includes a confidence score. Low-confidence regions are flagged so you know exactly where to verify.
Auto-detect source
Don't know what language is in the image? Leave source on auto — the pipeline figures it out.