Automatic mode performs a second pass for small or scattered text in screenshots, labels, receipts, and scanned pages.
Images
Image to Text (OCR)
Turn screenshots, receipts, scanned documents, and photos into editable text. Review the result, then copy it or download a text file.
Quick guide
How to use this tool
Add a clear JPG, PNG, WebP, or BMP image containing the text you want to capture.
Choose the text language and page layout, then select Extract text. The first use may take longer while the OCR language data is downloaded.
Review and correct the editable result, then copy it or download it as a plain-text file.
How to use this tool
Add a clear JPG, PNG, WebP, or BMP image containing the text you want to capture.
Choose the text language and page layout, then select Extract text. The first use may take longer while the OCR language data is downloaded.
Review and correct the editable result, then copy it or download it as a plain-text file.
Automatic checks the original plus an enlarged, high-contrast copy and keeps the stronger result. Use a specific layout when you know the image structure.
Source imageJPG, PNG, WebP, or BMP · up to 20 MB
Extracted textReady to extract text
From image to editable text
Capture text without retyping it.
Automatic mode checks both the original image and an enlarged, high-contrast copy, then keeps the stronger result. A tight crop and sharp, upright text still make the biggest difference.
Choose English, Filipino, Spanish, French, German, Italian, Portuguese, Japanese, Korean, or Simplified Chinese before scanning.
Recognition happens in the browser. The image and extracted text are not sent to ConduitTools or stored after you leave the page.
Frequently asked questions
Is my image uploaded?
No. The OCR engine processes the image inside your browser. It downloads its software and language data when needed, but your image and extracted text are not uploaded to ConduitTools.
Which image formats are supported?
You can use JPG or JPEG, PNG, WebP, and BMP images up to 20 MB. For the best result, use a clear image where the text is upright and fills most of the frame.
Can it read handwriting?
It is designed mainly for printed or screen text. Neat handwriting may produce a partial result, but cursive or faint handwriting is likely to need substantial correction.
Why are some words or numbers wrong?
OCR accuracy drops with blur, glare, tiny text, decorative fonts, tilted pages, and low contrast. Crop away unrelated graphics, choose the correct language and layout, and proofread important names, dates, totals, and URLs. The confidence value describes recognized text; it cannot tell you whether faint text was missed.
Does it preserve the original formatting?
It preserves line breaks and spacing where possible, but it does not recreate the original document design, tables, fonts, or columns exactly. The result is editable plain text.