Image to Text Converter
Pull the words out of a photo, a scanned page or a screenshot and copy them as editable text. The OCR runs in your browser, so your images never leave your device.
Images are read in your browser and never uploaded. Free, no sign-up, no watermark.
Drop images or screenshots here
JPG, PNG, WebP, HEIC, AVIF, GIF, BMP, TIFF · up to 20 images
Copied a screenshot? Press Ctrl+V (⌘+V on Mac) to paste it here.
Options
The English language model (2.9 MB) and the OCR engine (1.1 MB) download once from a public CDN and are then cached by your browser. Only these model files are downloaded; your files are read on your device and never uploaded.
Photos still called IMG_4821.jpg?
File Renamer AI names photos by what's in them, in bulk. 25 free renames.
How to use it
- Drop up to 20 images onto the page, click Choose images, or paste a screenshot with Ctrl+V (Cmd+V on Mac). JPG, PNG, WebP, HEIC, AVIF, GIF, BMP and TIFF all work.
- Pick the language of the text. English is the default; the model for a new language downloads once (0.7 to 3 MB) and is then cached.
- Click Extract text and watch each image being read. Turn on Black & white clean-up for faint text or photographed pages.
- Fix any misread words in the text box, then copy it or download everything as one .txt file. Untick Keep line breaks to join each paragraph into a single line.
About this tool
Image to text, or OCR (optical character recognition), is for the moments when the words you need are trapped in a picture: a screenshot of an error message, a photo of a whiteboard or a menu, a JPG of a receipt, a scanned letter, a slide someone posted as an image. Instead of retyping it, you get text you can paste into an email, a document, a translator or a search box.
The tool uses Tesseract, a long-standing open-source OCR engine, compiled to run inside your browser. Before reading, each image is decoded (HEIC and AVIF included), flattened onto white, scaled so its long side is at most 3000 pixels, and small screenshots are enlarged so their text is big enough to recognise. Clean printed or on-screen text in the right language usually comes out nearly perfect.
Accuracy drops with the quality of the picture. Blurry, low-resolution, tilted or curved pages, decorative fonts, text over busy backgrounds and handwriting all cause mistakes, and handwriting is often not read at all. Tables come out as lines of text without their columns. Check the confidence figure next to each result, and the text itself, before relying on it for anything important.
Good to know
- Works best on clear printed or on-screen text at least 10 pixels tall. Handwriting is read poorly or not at all.
- Blurry, low-resolution, tilted or curved photos lower accuracy. Crop to the text and shoot the page straight on for the best results.
- Tables and multi-column layouts come out as plain lines of text; columns and cell borders are not kept.
- Only the OCR engine (about 1.1 MB) and the language model download, from a public CDN, the first time. Your images are never uploaded.
Other ways to do it
iPhone or iPad (Live Text)
- Open the photo in Photos, or point the Camera at the text. Live Text needs iOS 15 or later on iPhone XS, XR or newer.
- Tap the Live Text button in the lower-right corner to highlight all the text, or touch and hold a word and drag the grab points.
- Tap Copy (or Select All, then Copy) and paste it anywhere. If nothing is detected, turn on Live Text in Settings > General > Language & Region.
Mac (Preview)
- Open the image in Preview (macOS Monterey or later) and hold the pointer over the text.
- Drag to select the text, Control-click the selection and choose Copy Text.
- If you cannot select anything, choose Tools > Text Selection and try again.
Windows (Snipping Tool)
- Open the image so the text is on screen, press Windows logo key + Shift + S and drag over the text.
- Open the snip in Snipping Tool by clicking the notification.
- Select Text actions, then Copy all text, or select just the lines you need and copy them. The recognition runs on your PC.
Windows (PowerToys Text Extractor)
- Install Microsoft PowerToys and turn on Text Extractor in its settings.
- Press Win+Shift+T and drag a box around the text on screen; it is copied to the clipboard.
- It reads only languages whose Windows OCR language pack is installed. Microsoft now suggests Snipping Tool for most uses.
Chrome (Google Lens)
- Right-click the image in Chrome and choose Search with Google Lens. Google must be your default search engine.
- In the side panel, select the text and copy it.
- This sends the image to Google, so avoid it for private documents.
Google Drive
- Upload the image (JPG, PNG or GIF, 2 MB or smaller) to Google Drive.
- Right-click it and choose Open with > Google Docs.
- A new Google Doc opens with the recognised text. Tables and columns are usually not kept.
Private by design: nothing is uploaded
Most online OCR sites upload your picture to a server to read it. Here the OCR engine runs inside your browser: the only downloads are the engine and the language model (public files, fetched once and cached), and your images never leave your device. That makes it safe for IDs, receipts, contracts and screenshots of private chats.
Step-by-step guides
Related tools
Frequently asked questions
How do I convert an image to text for free?
Drop the image onto this page (or paste a screenshot with Ctrl+V), choose the language and click Extract text. The text appears in a box under the image, ready to copy or download as a .txt file. It is free with no sign-up, and up to 20 images can be read at once.
Can I convert a JPG to text?
Yes. JPG is the most common input, along with PNG screenshots and iPhone HEIC photos. WebP, AVIF, GIF, BMP and TIFF work too. Multi-page TIFFs are read from the first page; for scanned multi-page documents, save them as a PDF and use OCR PDF.
How accurate is it?
For sharp printed or on-screen text in the selected language, the result is usually close to perfect. Accuracy falls with blur, low resolution, skewed or curved pages, stylised fonts and busy backgrounds, and handwriting is rarely read well. Each result shows a confidence score; below 60 percent, check the text carefully or try Black & white clean-up.
Are my images uploaded to a server?
No. The OCR engine runs inside your browser. The only downloads are the engine and the language model, which are public files fetched once from a CDN and then cached. Your images and the extracted text stay on your device.
Which languages are supported?
English, Spanish, French, German, Portuguese, Italian, Dutch, Hindi, Indonesian, Chinese (Simplified), Japanese, Korean and Arabic. Choose the language of the text before reading: using the wrong one makes accents and non-Latin scripts come out garbled.
Can it read handwriting?
Only very neat block capitals, and even then with mistakes. Tesseract is trained on printed text. For handwritten notes, Live Text on an iPhone or Google Lens tend to do better, though they are not perfect either.