How to Extract Text From an Image or Screenshot (Free OCR)

To get text out of a photo or screenshot, run it through OCR (optical character recognition), which reads the letters in the picture and gives you text you can copy, edit and search. On a computer or phone browser, upload the image to an OCR tool and copy the result. On a phone, Google Lens or iPhone Live Text can also do it from the camera or gallery.

The free image to text converter runs OCR inside your browser, so the image is not uploaded anywhere. It reads English only. This guide explains how OCR works, exactly what the tool does, how to get cleaner results, and which free options to use for Hindi and other Indian languages.

How OCR works, briefly

An OCR engine doesn't "see" words the way you do. It works in stages:

1. Clean-up: the image is turned into black and white, so each pixel is either text or background. This is called binarisation, and uneven lighting or shadows make it go wrong.
2. Layout: the engine finds blocks of text, then lines, then words.
3. Recognition: modern engines such as Tesseract (version 4 onwards) read a whole line at a time with a neural network trained on one language's script and common words.
4. Output: the recognised characters come out as plain text. Bold, colours, tables and columns are usually lost.

Each stage depends on the one before. A tilted line breaks the layout step, and a blurry letter gets guessed. That is why most accuracy tips are about the image, not the software.

Using the free image to text converter

1. Drop a photo, screenshot or scan on the box, or tap to choose a file (JPG, PNG or WebP work best).
2. Press Extract text. A progress bar shows the percentage while it reads.
3. The text appears in an editable box. Fix any mistakes there, then press Copy text.

What happens behind the scenes, from the tool's code:

• It uses Tesseract.js version 5, the JavaScript build of the open-source Tesseract engine. On first use your browser downloads the engine and the English language data from public servers; after that they are cached and later runs start faster.
• Recognition runs in your browser. The image itself is never sent to EthicalFusion or anyone else.
• It loads only the English model. Tesseract has Hindi and other Indian language models, but this tool doesn't use them, so Devanagari or other Indian scripts come out as nonsense or nothing. Numbers and common symbols in English text are read normally.
• If it finds no text, it says "No text detected" and suggests a clearer, larger image.

It takes images only. For a PDF, first check whether it already has real text: the guide to extracting text from a PDF covers that. For a scanned PDF, turn each page into an image with PDF to JPG and run OCR on the pages. iPhone photos saved as HEIC may not open, so convert them with HEIC to JPG first.

Example 1: an error message on a screenshot

You have a screenshot of an error dialog that won't let you select its text. Crop the screenshot to just the dialog with your phone or computer's built-in editor, upload it and press Extract text. A sharp screenshot of standard dark-on-light text usually comes out almost word for word; check long numbers and codes character by character, because one wrong digit in an error code sends a search in the wrong direction.

Example 2: a photographed notice

You photographed a printed notice at an angle and in poor light. Retake it first if you can: stand straight in front of it, fill the frame with the text, and avoid your own shadow. If the photo came out sideways, turn it with the image rotator, which turns it 90° clockwise. For an upside-down photo, download the result, upload it again and rotate once more. Then upload it to the OCR tool.

Tips for better accuracy

Resolution

The Tesseract documentation says it works best on images of at least 300 DPI, roughly what a document scanner produces. Google's own OCR help asks for text at least 10 pixels high. In practice: take the photo close enough that the text fills the frame, and screenshot at normal zoom rather than zoomed out. If a screenshot's text is tiny, enlarging it can help: in the image resizer, double both the width and the height (it has no aspect-ratio lock, so change both), then run OCR on the larger copy.

Contrast and lighting

Dark text on a plain light background works best. Tesseract's documentation recommends dark text on a light background for version 4 and later, so for screenshots of apps in dark mode, switch the app to its light theme before taking the screenshot. In photos, use even daylight, turn off the flash on glossy paper, and keep shadows off the text.

Skew and rotation

Tesseract's documentation says line detection gets much worse when a page is skewed. Hold the phone parallel to the page. Fix sideways or upside-down images with the image rotator; for a slight tilt, use the straighten or crop tool in your phone's photo editor, because the rotator only turns in 90° steps.

Cropping and clutter

Crop away everything that isn't the text you want: other columns, photos, logos and dark scanner borders. A tight crop gives the layout step less to get wrong.

Check the usual mix-ups

OCR commonly confuses 0 and O, 1 with l and I, 5 and S, and "rn" with "m". Check amounts, dates, PAN or account numbers, email addresses and phone numbers by eye before you use them.

Other free ways to copy text from images

• Google Lens reads Hindi, Bengali, Telugu, Marathi and Tamil text, according to Google India's July 2020 announcement. In Google Photos, open a photo, tap Lens, and select the words to copy.
• Google Docs converts JPG, PNG and GIF images and PDFs of up to 2 MB into an editable document. Its supported languages include Hindi, Marathi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam and Punjabi (Gurmukhi). The image must be the right way up, and lists, tables and columns usually aren't kept.
• Snipping Tool added Text actions in version 11.2308.33.0 in 2023. Select part of the text and press Ctrl+C, or copy it all at once. Text actions also has a Quick redact option that hides email addresses and phone numbers before you share a screenshot.
• Live Text is switched on under Settings, General, Language and Region. Apple's list of Live Text languages includes English (India) but not Hindi.

Privacy

Think about where the image goes before you OCR an Aadhaar card, PAN card, bank statement or contract. With the browser tool here, the image stays on your device; only the reading engine is downloaded. Google Docs works by storing the file in your Google Drive, so delete the image and the converted document afterwards if you don't need them. Whatever you use, be careful where you paste the extracted text: an ID number copied into a random website is as exposed as the photo.

Cleaning up the extracted text

OCR keeps the original line endings, so a paragraph from a book comes out as many short lines. The remove line breaks tool joins every line into one paragraph; run it one paragraph at a time if you want to keep paragraph breaks. For tidying stray spaces and other copy-paste mess, see the guide to cleaning up pasted text.

Frequently asked questions

Is my image uploaded when I use the image to text converter?

No. The OCR runs in your browser. The only download is the reading engine and English language data, the first time you use it.

Can it read Hindi?

No. The tool uses Tesseract's English model only. For Hindi and other Indian languages, use Google Lens or open the image with Google Docs in Google Drive.

Can OCR read handwriting?

Printed text works far better. Neat block capitals sometimes come out, but joined-up handwriting usually doesn't with Tesseract.

Why is my extracted text full of mistakes?

Usually the text is too small, blurry, tilted or in shadow, or it is light text on a dark background. Retake or crop the image, straighten it, and try again.

How do I copy text from a scanned PDF?

Convert each page to an image with PDF to JPG, then run the images through OCR. If the PDF already has selectable text, use a PDF text extractor instead; it is faster and exact.

Sources: Tesseract documentation, "Improving the quality of the output"; Google Drive Help, "Convert PDF and photo files to text"; Google India blog, July 2020; Windows Insider Blog, 14 September 2023 (Snipping Tool Text Actions); Apple Support, "Copy and translate text from photos on your iPhone or iPad", and Apple iOS feature availability page (checked October 2026).