How to Capture Text from Screen on Windows (Free)
Press one hotkey, drag a box around the words you see on your screen, and get them as editable text — no retyping, no screenshot app, no cost.
You cannot copy subtitles from a video, text burned into a scanned PDF, an error message in a dialog box, or a line in a game. That is exactly what “capture text from screen” means: point at any region of your display and have OCR turn the pixels into text you can paste anywhere. On Windows, the free way to do it is Capture2Text, a small portable utility built on the Tesseract OCR engine.
Step 1: Get Capture2Text running
There is no installer. Download the portable zip from the download page, extract it anywhere — a folder, your desktop, even a USB stick — and double-click Capture2Text.exe. The program sits quietly in the system tray. It is Windows-only (no macOS or mobile version exists) and completely free, licensed under GPLv3. Version v4.6.3 is the latest release (2022-03-19), and it remains fully functional on Windows 10 and 11.
Step 2: Press Win+Q and select the text
- Press Win+Q (the default “capture line” hotkey).
- Your cursor becomes a crosshair. Click and drag a rectangle around the words you want — a subtitle, a table in a scanned document, a line of code in a screenshot.
- Release the mouse. Capture2Text runs OCR on that region and processes the result according to your output settings.
That is the whole workflow. If you would rather grab a single word, vertical text (Japanese or Chinese columns), or have the result spoken aloud, hotkey setup explains the other modes. There is also a “Bubble OCR” mode that shows the recognized text in a speech-bubble-style tooltip right where you selected it.
Step 3: Choose where the text goes
By default the recognized text is copied to your clipboard, ready to paste. But you have options in the settings:
- Clipboard — the default. Capture and paste with Ctrl+V.
- Popup window — shows the text in a small window so you can check it before copying.
- Append to file — every capture is saved into a text file, handy when you are collecting quotes from many pages.
- Translate — sends the text to Google Translate and shows the result. This is the one feature that needs an internet connection; the OCR itself works fully offline.
- Speak — reads the text back using Windows text-to-speech.
Which languages can it read?
Seven languages work out of the box: English, French, German, Japanese, Korean, Russian, and Spanish. More than 90 languages are supported in total — extra language data can be added from the language packs section on the download page. (Italian is not among the built-in set; you would need to add it as an extra pack.) For tips on better results, see how to improve OCR accuracy.
What “capture text from screen” is good for
- Copying subtitles or lyrics from videos
- Extracting text from scanned PDFs or photos where the text is not selectable
- Copying error messages or dialog text from programs
- Grabbing a word or sentence from a game for quick translation
- Reading text inside images shared in chat apps
FAQ
Is capturing text from screen really free?
Yes. Capture2Text is free and open-source (GPLv3). There is no trial, no paid tier, and no feature locked behind a paywall. On SourceForge it holds a rating of 4.7/5 from 89 user votes.
Does it work on PDFs?
It works on anything visible on your screen — including a PDF open in your viewer. It reads the rendered pixels, so even scanned PDFs with non-selectable text are fine.
Does it upload my screenshots?
No. OCR runs entirely on your computer. The only network call is the optional Google Translate step, which you can leave disabled.
Is there a Mac version?
No — Capture2Text is Windows-only. There are dedicated Mac alternatives for screen OCR listed here.
Can it batch-process many images at once?
No. Capture2Text is designed for one quick selection at a time. If you need batch OCR of many image files, Umi-OCR or VietOCR are the free offline tools built for that.