image-to-image
alibaba/qwen-image/translate
Advanced AI model for image understanding, OCR, and text translation with Alibaba Qwen Vision.
Idle
Your request will cost $0.01 per run.
For $1 you can run this model approximately 100 times.
One more thing:
Examples
View all
README
Alibaba Qwen Vision - Image Understanding & Translation
Alibaba Qwen Vision is a multimodal AI model that integrates OCR (optical character recognition) and multilingual translation. Built on Alibaba Cloud’s DashScope, it can extract text from images and translate it into multiple languages quickly and accurately.
Why it looks great
- Accurate OCR: recognizes printed and handwritten text from images.
- Multi-language support: detect and translate across English, Chinese, Japanese, Korean, French, German, Spanish, Russian, Arabic, and more.
- Customizable translation: define terminologies and filter sensitive words for domain-specific use cases.
- Document understanding: works with forms, receipts, signage, and scanned documents.
- Real-time performance: fast turnaround for practical scenarios like menus, signs, and learning materials.
Limits and Performance
- Supported formats: PNG, JPEG, WEBP
- Processing speed: ~3–6 seconds per image
- Segmentation: automatic text region detection (can be disabled via
skip_image_segment)
Pricing
| Task Type | Cost per image |
|---|---|
| OCR / Translation | $0.01 |
How to Use
- Upload the image containing text.
- Select source_lang (e.g., auto, en, zh, ja, ko, fr, de, es, ru, ar).
- Select target_lang for translation.
- (Optional) Add sensitives to filter sensitive words.
- (Optional) Add terminologies to ensure consistent translations for key terms.
- (Optional) Check skip_image_segment if you don’t want automatic segmentation.
- Run the job and download/view the results.
Pro tips for best quality
- Upload high-resolution images with clear text for better OCR accuracy.
- Use auto for source_lang when handling mixed or unknown languages.
- Add terminologies for industry-specific vocabulary (e.g., finance, medicine).
- Filter sensitive words via sensitives for safer outputs.
- Keep segmentation enabled for documents with multiple text regions.
Notes
- Best for document digitization, translation of signage/menus, multilingual education, and accessibility tools.
- If you did not upload the image locally, please ensure that the image URL is accessible! A successfully accessible image will display a preview in the interface.