Extract printed English and Traditional Chinese text from a screenshot, document photo or image. Review and edit the result before sharing.
Images and recognized text stay on your device. First use downloads the OCR engine and your selected language data from this website; the page still makes network requests for these resources.
1. Upload Image
You can also paste an image with Ctrl+V / ⌘V when browser permissions allow. Pasting into the result editor keeps normal text editing.
Image preview
Adjustments affect the OCR copy only. Your original file is unchanged. EXIF orientation is applied by the browser; use Rotate for sideways text.
Standard limit: 20 MiB, 24 megapixels and 12,000 px per side. Mobile or devices reporting ≤4 GB RAM: 8 MiB, up to 12 megapixels and 8,000 px per side. The source cap falls to 6 MP on ≤2 GB RAM or mobile devices without a memory hint. OCR images are reduced to 4 MP (2 MP on these devices).
Downloads only the chosen data: English ≈2.8 MiB; Traditional Chinese ≈1.6 MiB; mixed uses both. Engine download is approximately 4 MiB, depending on browser capabilities. Switching languages updates the local recognizer.
This is the recognizer’s estimate, not a measured accuracy guarantee. Proofread names, numbers and punctuation.
How to extract text from an image
01
Choose one JPG, PNG or WebP image, or paste a screenshot from your clipboard.
02
Select English, Traditional Chinese or both. Rotate sideways text; try grayscale or contrast enhancement when useful.
03
Extract text and wait for local recognition. The first run downloads the engine and selected language data.
04
Proofread and edit the result, then copy it or download a UTF-8 TXT file.
Screenshots, JPG photos and PNG documents
Clear printed text works best. Crop out unrelated content, keep the document flat and well lit, and avoid reflections. Large images are reduced to protect memory; crop a dense page into a smaller region if tiny characters are lost. The tool keeps useful line breaks where recognized, but does not recreate the original layout.
English and Traditional Chinese OCR
Choose the language that matches the text. Mixed English + Traditional Chinese loads both models and can use more memory. Simplified Chinese is not a separate model in this release. Recognition quality depends on font, resolution, contrast and layout; check Chinese characters and punctuation carefully.
Image OCR and PDF text extraction serve different files
PDF to Text reads a PDF’s existing selectable text layer; it does not recognize scanned pixels. Use that tool for a digital PDF. For a scan, export a page as an image with PDF to Image, then recognize that image here. Image to Text does not create searchable PDFs or edit the source image.
Choose a JPG, PNG or WebP image, select its OCR language, and click Extract Text. Review the local recognition result, edit mistakes, then copy or download the text.
Can I convert a screenshot into text?
Yes. Upload a screenshot or paste an image when your browser supports clipboard images. Sharp text with good contrast is easier to recognize than a blurred photo or a complex layout.
Does online OCR support Traditional Chinese?
This tool provides a Traditional Chinese model and an English + Traditional Chinese option. Models download only when needed. Output may contain mistakes and needs proofreading.
Can OCR recognize handwritten text?
These models are intended mainly for printed text. Handwriting, unusual fonts, blurred photos and complex layouts may produce inaccurate or empty results; handwriting recognition is not guaranteed.
Is my image uploaded to a server?
No. Image decoding and OCR run locally in browser workers. The website downloads its OCR engine and selected language data, but does not send your image or recognized text to an OCR service or server.
What is the difference between OCR and PDF text extraction?
OCR recognizes characters from image pixels. PDF to Text extracts an existing text layer without recognizing pixels. A scanned PDF without a text layer needs OCR; export a page to an image first for this tool.
