OCR — optical character recognition — converts images of text into machine-readable text you can edit, copy, and search. Common uses: extracting text from a scanned PDF so you can search it, copying text from a screenshot or photo, digitizing printed documents, or making a PDF searchable when the original was created by scanning paper. The quality of free tools has improved dramatically; for clean printed text, free OCR is often indistinguishable from paid solutions.
Tools You May Already Have
Microsoft OneNote (Free with Windows / Microsoft Account)
OneNote includes a hidden but capable OCR feature. To use it:
- Open OneNote (Desktop app, not the web version).
- Insert the image: Insert → Pictures, or paste the image directly onto a OneNote page.
- Right-click the image and select "Copy Text from Picture."
- Paste the extracted text anywhere.
OneNote's OCR handles clean, printed text well and supports multiple languages. It processes the image locally on your computer.
Google Drive / Google Docs (OCR via Upload)
Google Drive can OCR images and PDFs during upload:
- Upload a PNG, JPG, or PDF to Google Drive.
- Right-click the uploaded file and choose "Open with → Google Docs."
- Google Docs opens the file with the image at the top and extracted text below it.
- Copy the text section.
This works well for clean printed documents. Handwriting recognition is inconsistent. Note that files are uploaded to Google's servers.
Windows 11 PowerToys — Text Extractor
Microsoft's free PowerToys utility package includes Text Extractor, which uses OCR to let you select any region of your screen and copy the text from it. Press the configurable hotkey (default Win+Shift+T), draw a selection box over any text visible on screen — in a PDF, an image, a video frame — and the text is copied to your clipboard. This works on any content that is visually on screen, not just files. Download PowerToys free from the Microsoft Store or GitHub.
Free Desktop Software: Tesseract-Based Tools
Tesseract is an open-source OCR engine originally developed by HP and currently maintained by Google. It underpins many free OCR tools and is among the most accurate open-source OCR engines available for printed Latin-script text.
fre:ac / NAPS2 (Free, Windows)
NAPS2 (Not Another PDF Scanner 2, naps2.com) is a free, open-source document scanning and OCR application for Windows. Key features:
- Scan directly from a scanner or import existing images and PDFs
- OCR via Tesseract to create searchable PDFs — the document looks the same visually, but the text layer makes it fully searchable and copy-pasteable
- Supports over 100 languages via downloadable Tesseract language packs
- Batch processing for multiple documents
- All processing happens locally — nothing uploaded
NAPS2 is particularly useful for creating searchable PDFs from scanned documents, which is different from extracting raw text — the output is still a PDF but with a hidden text layer that makes it searchable in any PDF viewer.
Online Tools (When Uploading Is Acceptable)
OnlineOCR.net (Free, No Account Required)
OnlineOCR.net converts uploaded images or PDFs to editable text, Word, or Excel format. The free tier allows five guest conversions per hour without an account. Output formats include plain text, Word (preserving some formatting), and Excel (useful for tables). Accuracy on clean documents is good. Files are uploaded to their servers.
Adobe Acrobat Online (Free Tier)
Adobe offers free online PDF tools including OCR at acrobat.adobe.com. The free tier allows limited conversions per month. The output is a searchable PDF. Adobe's OCR accuracy is among the best available, particularly for complex layouts with multiple columns. Adobe account required.
Tips for Better OCR Results
OCR accuracy depends heavily on the quality of the source image. Practical steps to improve results:
- Resolution: 300 DPI (dots per inch) is the standard recommendation for OCR. Images scanned below 150 DPI will have significantly more errors. When scanning, set the scanner to at least 300 DPI in grayscale or color.
- Deskew before OCR: Tilted images produce many errors. NAPS2 has an automatic deskew option. Even a 2–3 degree tilt significantly reduces accuracy.
- Clean backgrounds: Coffee stains, creases, and bleed-through from the other side of thin paper all reduce accuracy. A contrast-boosting step in an image editor before OCR can help.
- Font matters: Clean, high-contrast, standard fonts (Times, Arial, Courier) are recognized at very high accuracy. Decorative fonts, handwriting, and stylized logos will have high error rates regardless of tool.
- Check the output: OCR is not 100% accurate even on good source material. Numbers and similar characters (0/O, 1/l/I, 5/S) are the most common misread characters. Always proofread extracted text for anything where exact accuracy matters.
Choosing the Right Tool
- One-off extract from an image on screen: Windows PowerToys Text Extractor
- Extract text from an uploaded image or PDF: Google Drive OCR or OneNote
- Create searchable PDFs from scans (offline): NAPS2
- High-accuracy online OCR, convert to Word: OnlineOCR.net or Adobe Acrobat online
- Batch processing, many documents, offline: NAPS2 with Tesseract