How to Extract Text from Screenshots and Photos — Free OCR Guide
You have a screenshot of a code snippet from a tutorial video. Or a photo of a whiteboard from a meeting. Or a scanned contract that exists only as an image in a PDF. The text is right there — you can read it with your eyes — but you cannot select it, copy it, or search for a specific word. This is exactly the problem OCR solves, and it has gotten remarkably good in the past few years.
What Is OCR and How Does It Work?
OCR stands for Optical Character Recognition. It is a technology that analyses the pixels in an image, identifies patterns that correspond to letters, numbers, and symbols, and converts those patterns into actual editable text. Modern OCR goes beyond simple pattern matching — it uses machine learning models that understand context, can handle multiple fonts and sizes in the same image, and even recognise text at slight angles or on curved surfaces.
The iFormat OCR tool processes your images directly in the browser. Upload a screenshot, photo, or scanned document, and it extracts all the text it can identify. You can then copy the extracted text, edit it, or save it for use in documents, spreadsheets, or emails.
When Do You Need OCR?
The most common scenarios are more everyday than you might think. Screenshots: You took a screenshot of an error message, a recipe, a social media post, or a code block, and now you want the text without retyping it. Scanned documents: Old contracts, receipts, or forms that were scanned to PDF as images — the text exists visually but is not selectable. Whiteboard photos: Meeting notes captured on a phone camera that you need to transcribe into meeting minutes.
Handwritten notes: Lecture notes, journal entries, or planning sketches photographed for digital archiving. Business cards: Rather than manually typing someone's contact details, photograph the card and extract the text. Book pages and articles: Extracting quotes or data from printed material without retyping entire paragraphs. Receipts and invoices: Pulling amounts, dates, and vendor names from photographed receipts for expense tracking.
Tips for Better OCR Accuracy
Maximise OCR Accuracy
Resolution matters: Higher-resolution images produce better results. If scanning a document, use at least 300 DPI.
Contrast is key: Dark text on a light background works best. Avoid photos with shadows falling across the text.
Straight alignment: Text that is level and square to the camera is easier for OCR to process. Skewed or rotated text reduces accuracy.
Avoid compression artefacts: Heavily compressed JPEGs blur the edges of letters. Use PNG for screenshots or high-quality JPEG for photos.
If your image is dark or low-contrast, consider adjusting the brightness and contrast before running OCR. A quick adjustment in your phone's built-in photo editor — increasing brightness and contrast slightly — can dramatically improve text recognition accuracy, especially for whiteboard photos taken in dim conference rooms.
Multi-Language Support
Modern OCR engines support dozens of languages, including Latin-alphabet languages (English, Spanish, French, German), Cyrillic (Russian, Ukrainian), Arabic, Chinese (simplified and traditional), Japanese, Korean, Hindi, and many more. The quality varies by language — Latin-alphabet languages tend to have the highest accuracy because they have the most training data, but CJK (Chinese, Japanese, Korean) recognition has improved dramatically in recent years.
If your document contains multiple languages (common in academic papers or international contracts), OCR can typically handle the mixed text, though accuracy may dip at the boundaries between scripts. For best results with non-Latin scripts, make sure the image is high-resolution and the text is clearly printed rather than handwritten.
What OCR Cannot Do (Yet)
OCR has limitations worth understanding so your expectations match reality. Handwriting recognition is still inconsistent — neat, printed handwriting works reasonably well, but cursive or messy handwriting produces unreliable results. Decorative and stylised fonts (the kind used in logos, posters, and artistic designs) often confuse OCR engines because the letter shapes deviate significantly from standard typefaces.
Very small text, text overlaid on busy backgrounds (like text on a photograph), and heavily degraded or faded text all reduce accuracy. OCR also does not preserve the original formatting — it extracts raw text, not the layout with columns, tables, and indentation. For structured data like tables, you may need to reorganise the extracted text manually. Despite these limitations, for standard printed text in clear images, modern OCR accuracy is above 99%, which makes it faster and more reliable than manual transcription.
Once you have extracted the text, you might want to put it into a clean document format. You can paste it into a word processor and convert to PDF, or if the original image is in an unusual format, convert it to a standard format before running OCR for the best results.
