Convert photos, screenshots, and scans to editable text with free online OCR.
Image OCR
Convert photos, screenshots, and scans to editable text with free online OCR.
How to turn images into editable text
Convert photos, screenshots, and scanned pages
Upload a receipt, notice, screenshot, menu, or scanned page when you need the words as editable text. OCR creates a draft you can copy, quote, or reuse instead of retyping it.
Choose the recognition language
The OCR engine loads the selected language model in the browser before reading the image. Matching the language to the text in the picture helps names, accents, kana, or Chinese characters produce cleaner output.
Prepare the image, then proofread the OCR result
Rotate sideways text or switch between original colors, grayscale, and high contrast before recognition. The review stage flags lower-confidence areas for attention, but it does not auto-correct them; edit the draft line by line before copying or downloading it.
FAQ
Common questions and answers about this topic.
How do I turn an image into text?
Select or drop in an image, choose the recognition language, and the tool reads the picture inside your browser. The result opens as a proofread draft you can edit line by line before you copy it, download it as a .txt file, or share it. To count words or characters afterwards, paste the text into Text Statistics.
Which image formats and sizes are supported?
JPG, PNG, WebP, BMP, TIFF, and GIF are accepted, one image per run. A file can be up to 10 MB, and the picture itself up to 25 megapixels; anything larger is refused with a message showing its actual pixel dimensions.
Which recognition language should I choose, and what about mixed English and Chinese?
Pick the one that matches the text in the picture: English, Traditional Chinese, Simplified Chinese, Japanese, Korean, Spanish, or French. For pages that mix English with Chinese, choose English + Traditional Chinese or English + Simplified Chinese so both scripts are read in a single pass. The language model is loaded into the browser before recognition starts, so the first run in a new language takes longer.
Can it read handwriting?
Not reliably. The recognition models are trained on printed and typeset characters, so printed documents, screenshots, signs, receipts, and book pages give the best results. Handwritten notes, especially cursive or uneven writing, often come back with missing or wrong characters, and no preprocessing option changes that.
Is the image uploaded to a server?
No. Recognition runs entirely in your browser, so the picture never leaves your device and is not sent anywhere. The only thing fetched from the network is the language model, and that traffic travels towards your browser rather than away from it.
Can I run a PDF through it?
No, this tool accepts images only. If the PDF already carries a text layer, PDF to Text pulls that text out directly without OCR. For a scanned or image-only PDF, export or screenshot the page as an image first and recognize it here.
The result has mistakes — how do I improve it?
Check the recognition language first, then rotate sideways pages and switch the image between original colors, grayscale, and high contrast before running it again; sharp, high-contrast text reads best. The review stage marks the areas it is least confident about and shows a confidence score, but it never corrects them for you, so edit the draft line by line. Layout is not rebuilt either: columns, tables, and images are returned as plain text.