Image to Text Converter (OCR)
AIConvert an image to text you can copy, search and edit. This free image to text converter reads screenshots, scanned pages, receipts, slides and photos of documents with Tesseract OCR running in your browser, in 21 languages. Copy the text or download it, with word positions if you need them. Your images are never uploaded.
- Runs on your device
- On-device AI
- Works offline
Updated
Drop your images here
How to extract text from an image
Step 1: Add the pictures that hold the text, such as screenshots, scans or phone photos in JPG, PNG, WebP or HEIC. Several at once is fine.
Step 2: Choose the Text language. Its language pack downloads once, after you confirm.
Step 3: Keep Plain text, or pick Text + positions (JSON) for word coordinates.
Step 4: Press Extract, then copy the text or download it as a file.
Good to know
Settings explained
- Text language
- A small language pack downloads the first time you use each language.
- Auto-enhance before reading
- Enlarges small text, boosts contrast, inverts dark-mode screenshots and straightens slightly tilted scans.
Formats
Batch-friendly: add many files and they are processed one after another — if one fails, the rest still finish.
How it runs
- Vision model
- A compact computer-vision model runs locally in your browser. Nothing is sent to an AI service.
What image to text reads well, and what it does not
OCR is very accurate on clean printed text and screenshots, where letters are sharp and evenly lit. It gets harder as the text gets smaller, blurrier, more decorative or handwritten. The result card shows a confidence score, so you can tell at a glance whether to double-check.
| Image | What to expect | Tip |
|---|---|---|
| Screenshots, including dark mode | Near-perfect text | Auto-enhance inverts dark screenshots for you |
| Scanned documents and printed pages | Very accurate | Scan straight; slight tilts are corrected |
| Phone photos of pages, slides, signs | Good when flat, sharp and well lit | Crop to the text before extracting |
| Receipts | Good; faded thermal print is harder | Photograph in daylight, flat on a table |
| Handwriting | Unreliable | Neat block capitals work best |
| Logos, stylised or curved text | Often wrong | Type short text by hand |
What is OCR?
OCR (optical character recognition) is software that finds the letters in an image and turns them into real, editable text. Modern OCR engines such as Tesseract, which this page uses, read whole lines with a neural network rather than matching letters one by one, which copes far better with different fonts.
Before reading, Auto-enhance prepares the image: small images are enlarged so letters are big enough, colours become high-contrast greyscale, dark-mode screenshots are inverted, and slightly tilted scans are straightened. Turn it off under Advanced settings only if an already perfect scan comes out worse.
Copy text from a screenshot
Screenshot to text is the easiest case for OCR, because the letters were drawn by a computer. Take the screenshot as usual, paste it straight onto this page with Ctrl + V (⌘ + V on a Mac) or drop the saved file, and press Extract. Error messages, chat conversations, code on a slide and text in an app that blocks copying all come out as plain text. Line breaks follow the layout, and runs of empty lines are tidied up.
Which languages can it read?
Twenty-one languages, each with its own small language pack that downloads the first time you use it and then stays in your browser. Choose the language the text is written in; for a page that mixes languages, pick the main one.
| Script | Languages | Example pack size |
|---|---|---|
| Latin | English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Turkish, Indonesian, Vietnamese | English 3.0 MB, French 0.7 MB, Dutch 3.0 MB |
| Cyrillic | Russian, Ukrainian | — |
| Arabic script | Arabic, Urdu | — |
| Indic | Hindi, Bengali | Hindi 1.4 MB |
| East Asian | Chinese (Simplified and Traditional), Japanese, Korean | Chinese (Simplified) 1.7 MB |
JPG to text, PNG to text and other formats
Any image your browser can open works: JPG, PNG, WebP, HEIC from iPhones, AVIF, BMP, TIFF and more. Drop several and each gets its own result. Choose the output that suits the next step.
| Output | What you get | Use it for |
|---|---|---|
| Plain text | The text with its line breaks, as a .txt file or copied | Notes, emails, documents, translation |
| Text + positions (JSON) | Blocks, paragraphs, lines and words, each with a box and a confidence score | Developers, redaction, search indexes |
Image to text on iPhone, Android, Mac or Windows
Phones and computers have built-in text recognition for quick copies. This page adds batches, 21 languages, a confidence score, JSON positions and saved files.
iPhone and iPad
Live Text lets you select text in Photos and the camera. For a file, many images or another language, open this page in Safari.
Android
Google Lens copies text from the camera or gallery. Use this page in Chrome when you want a text file or do not want to use an online service.
Mac
Live Text works in Preview and Photos on recent macOS. Drag a folder of scans here to extract them all at once.
Windows 11
The Snipping Tool’s Text actions copies text from a screenshot. For scans and photos in bulk, drag them into Chrome or Edge here.
Get better results from a photo of text
- Fill the frame with the text. Hold the phone parallel to the page and close enough that small print is sharp.
- Light it evenly. Shadows across a page and glare on glossy paper hide letters.
- One language per run. Pick the main language; for a bilingual page, run it once per language and combine.
- Check the confidence. When the score on the result card is low, compare the text with the image before you rely on it.
Is image to text free and private?
- Free. No account, page limit or watermark.
- Private. IDs, bank letters, receipts and contracts are read on your device; nothing is uploaded.
- Offline once the engine and your language pack are cached.
- Batch. Extract text from a whole folder of images in one run.
Tips & troubleshooting
Crop to the text first with crop image; fewer distractions mean fewer stray characters.
Want the scans as one document instead? Put the images into a PDF with JPG to PDF.
Text from a video or a voice recording needs speech recognition: use audio to text.
Subtitle files are already text: SRT to TXT strips the timings without OCR.
Image to Text (OCR) FAQ
How do I extract text from an image?
Drop the image here, choose the language of the text and press Extract. Tesseract OCR reads it in your browser and shows the text, which you can copy or download as a .txt file. Several images can be processed in one go.
Can I extract text from an image for free?
Yes. There is no charge, sign-up or page limit, and no watermark. The OCR engine loads from this site and runs on your device; the only outside download is a small language pack, once, after you confirm it.
What is OCR?
OCR stands for optical character recognition: software that recognises letters in an image and turns them into editable text. It is what lets you copy text from a screenshot, a scanned page or a photo of a document.
How do I convert JPG to text?
Drop the JPG here and press Extract. Keep Auto-enhance on for phone photos: it enlarges small text, boosts contrast and straightens slight tilts before reading. Copy the result or download it as plain text.
How do I copy text from a screenshot?
Copy or save the screenshot, paste it here with Ctrl + V (⌘ + V on a Mac) or drop the file, and press Extract. Screenshots are the most accurate case for OCR, and dark-mode screenshots are inverted automatically. The text appears ready to copy, with its line breaks kept.
Can it read handwriting?
Only neat, printed-style handwriting, and not reliably. Tesseract is built for printed and on-screen text. For handwritten notes, expect to correct a fair amount, or type short notes by hand.
What gets downloaded, and are my images uploaded?
Your images are never uploaded. The OCR engine loads from this site, and each language pack downloads once from the jsDelivr CDN after you confirm. Both stay cached in your browser, so it works offline afterwards.