Tesseract LSTM
Image to text (OCR)
Recognize English and Persian text in an image, entirely on your device.
Model tools run on your device. Files and text are not uploaded.
Tesseract recognizes English and Persian. Files download on first use (roughly 5–15 MB depending on language/runtime). JPG, PNG and WebP: up to 8 MB and 6 megapixels. PDF: up to 10 MB, 50 source pages, and five selected pages per run. OCR can make mistakes; tables and complex layouts are not preserved.
Source text
Result
Choose an input, then run the tool.
Method, limits & licensing
Tesseract recognizes English and Persian. Files download on first use (roughly 5–15 MB depending on language/runtime). JPG, PNG and WebP: up to 8 MB and 6 megapixels. PDF: up to 10 MB, 50 source pages, and five selected pages per run. OCR can make mistakes; tables and complex layouts are not preserved.
Model sources, versions and checksumsA guide to this tool
How to use Image to text (OCR)
- Choose an image or PDF and select the text language.
- Choose the text layout and, for PDF, a range of up to five pages.
- Run the model, compare the text with the original, then copy or download.
A sample you can try
Sample input
Image: PixelDock OCR 2026 Local models read this image.
Sample output
PixelDock OCR 2026 Local models read this image.