A queue without the upload
Add up to 10 images or 100 MB, then stop, continue, retry, remove, and edit each result. One image is decoded at a time to avoid stacking full-size image memory.
Private, local OCR
Turn any image into editable text
Drop up to 10 screenshots, scans, or photos. BrowserOCR reads them one at a time on your device—nothing gets sent to a server.
✓ No uploads✓ English + Chinese✓ Works offline after caching
Recognition model
Preparing private OCR… Editable result
Choose a completed image to edit its result.
Two sizes, one private engine
Fast mode keeps the first model download compact. High accuracy is an optional recognition upgrade and is downloaded only after you choose it.
Built for multilingual text
The bundled PP-OCRv6 recognition dictionary supports English, simplified and traditional Chinese, and mixed lines.
Text, table Markdown, and ZIP
Edit each text or table Markdown result. Download one result or package every successful item into a locally generated ZIP.
Common questions
Frequently asked questions
Does BrowserOCR upload my images?
No. Image decoding, text detection, recognition, editing, and ZIP creation run locally in your browser.
Which image formats and batch limits work?
Choose, drop, or paste up to 10 JPEG, PNG, WebP, BMP, TIFF, HEIC, or HEIF images. Each image can be up to 50 MB and the queue can be up to 100 MB.
What is High accuracy mode?
It replaces the Tiny recognition model with a larger Small model. Switching modes clears prior results and reruns the current queue.
Can I use it offline?
Yes, after the app and the required OCR and image-decoder resources have been cached during an online visit.
Quick answer
What does BrowserOCR extract?
BrowserOCR extracts editable text and table Markdown from up to 10 images on your device. A queue can total 100 MB and is processed sequentially with PP-OCRv6; selected images and results are not uploaded.
| Image formats | JPEG, PNG, WebP, BMP, TIFF, HEIC, HEIF |
|---|---|
| Batch limits | 10 images, 50 MB each, 100 MB total; processed in order |
| Recognition | English, Chinese, and mixed lines |
| Output | Editable text and table Markdown, individual files, and a results ZIP |
| Processing | On this device; no image or recognized-text upload |
Table Markdown example
| Item | Total | | --- | --- | | Coffee | $4.50 |
Image OCR
Batch images to text
Queue photos and screenshots, edit each result, and export a ZIP.
Chinese OCR
Read mixed Chinese text
Recognize simplified, traditional, English, and mixed lines locally.
Zero upload
Keep sensitive images local
Understand the browser-only processing and network boundary.
Private document OCR
PDF to Text
Process one PDF at a time; PDF bytes, passwords, and results stay local.