Image-to-text OCR reads visible characters in an image and returns editable text. Upload a clear JPG, PNG or WebP, choose the recognition language, review the extracted words and download TXT or JSON.
Supported formats
Upload JPG, PNG or WebP images. Copy recognized text or export TXT and JSON. Images have no audio timeline, so image OCR does not produce SRT subtitles.
How it works
1. Choose a clear image
Upload a screenshot or photo with readable text. Crop unnecessary borders and use a straight, well-lit image. A higher-resolution source can help when characters are small.
2. Choose the text language
Choose Chinese, English, Japanese, Korean, Spanish, French or German. Recognition runs on the website server. Automatic selection uses the installed language models and does not guarantee correct script detection.
3. Review and save the words
Check names, numbers, punctuation and reading order. Edit the result, save changes, then copy the text or download TXT and JSON.
Limits and privacy
Images must be JPG, PNG or WebP, no larger than 5 MB and 20 megapixels. Results are accessible through the current browser session and are removed after 24 hours. No payment or subscription is required.
What to check before sharing
OCR returns text rather than an editable replica of the original layout. Handwriting, decorative fonts, curved surfaces and complex columns may be read incorrectly. This tool accepts individual images, not PDF documents or platform URLs.
Common questions
Can I extract text from a screenshot?
Yes. Save the screenshot as JPG, PNG or WebP, upload it and review the recognized text.
Why is SRT unavailable for images?
SRT requires start and end times. A still image has no speech timeline, so image extraction exports TXT and JSON instead.
Does OCR preserve a table or page layout?
No. The output is editable text. Check the reading order and reconstruct any table or complex layout yourself.