The best OCR tools to copy text from the screen in 2026

Some text cannot be selected: an error in a dialog box, a slide in a video call, a scanned page, a frame of a tutorial, a label inside an image. Screen OCR (optical character recognition) solves that. You drag a box over the text, and the tool reads the pixels and gives you text you can paste. The best OCR tools to copy text from the screen are free, fast and private, because the best ones read the text on your own computer instead of uploading your screen.

This guide compares eight tools for Windows and Linux (and a few that also run on a Mac) on speed, accuracy, languages, privacy and what happens to the capture afterwards, with prices as of September 2026.

The best OCR tools to copy text from the screen, at a glance

1. PowerToys Text Extractor: the fastest free way to copy text from anywhere on a Windows screen.
2. Barside: the best for keeping what you capture, with its text, in searchable notes, on Windows, Linux and Mac.
3. Snipping Tool: the best built-in option, with nothing to install on Windows 11.
4. NormCap: the best free, open-source tool for Windows, macOS and Linux, and it reads QR codes too.
5. Text Grab: the best for power users on Windows, with several capture modes and OCR engines.
6. ShareX: the best if you already use ShareX for screenshots.
7. Snagit: the best for editing the text inside a screenshot, not only copying it.
8. Click to Do: built into Windows 11 on Copilot+ PCs, for copying and acting on text with a click.

How we chose

We looked at tools that copy text from any part of the screen, and compared them on:

• Speed: how many steps from "I need that text" to text on your clipboard.
• Accuracy and languages: which OCR engine the tool uses and how you add languages.
• Privacy: whether the text is read on your computer or sent to a server.
• What you keep: only the text, or the image and the text together.
• Platforms and price: Windows, Linux and macOS, with prices from each product's official pages in September 2026.

Barside is our product. We include it where it fits, and say where others are better. For a one-off copy on Windows, PowerToys Text Extractor is quicker than Barside. Barside ranks second because it also runs on Linux and keeps each capture in your notes, where its text stays searchable.

Comparison table

How screen OCR works

Screen OCR tools do the same three things: they take a picture of the part of the screen you select, run an OCR engine over the picture, and hand you the text. The differences come from the engine and from what happens next.

• Windows OCR is the text recognition built into Windows. PowerToys Text Extractor, ShareX, Text Grab and Barside use it. It reads only the languages whose OCR language pack is installed, so a missing pack is the first thing to check when a tool finds no text on a page in another language.
• Tesseract is an open-source OCR engine. NormCap uses it on every system, Text Grab offers it as an option, and Barside uses it on Linux.
• Apple's Vision framework is the text recognition built into macOS, so there is nothing to set up. Barside for Mac uses it.
• ABBYY's engine powers Snagit's Grab Text.
• Click to Do on Copilot+ PCs runs OCR on your computer over a screenshot of the screen, and Microsoft says the analysis is always done locally.

Browser tools that search what you see work differently. Google Lens in Chrome, for example, sends a screenshot of the page and page data to Google when you use it on a page. That can be handy for a public web page, but it is the wrong choice for private documents, and it only sees what is in the browser tab, so we left it out of this list.

OCR is never perfect. Microsoft's own documentation for Text Extractor tells you to proofread the output, and that advice applies to every tool here. Numbers, code and unusual fonts deserve a second look.

The best OCR tools, reviewed

1. PowerToys Text Extractor

Text Extractor is part of Microsoft PowerToys, a free, open-source set of utilities for Windows. Press Win+Shift+T, drag over any text on the screen, including text inside images and videos, and the text lands on your clipboard. The capture closes by itself as soon as the text is copied. Hold Shift to move the selection instead of resizing it, and press Esc to cancel. Its code is based on Text Grab, which is also on this list.

Best for: anyone on Windows who wants text on the clipboard with the fewest steps.

Strengths: it is quick, free and runs on Windows 11 and Windows 10 version 2004 or later, on x64 and ARM64. You choose a preferred OCR language in its settings.

Limits: it keeps nothing but the text on your clipboard, with no image and no history. It can only read languages whose Windows OCR pack is installed, and the default language follows your Windows language settings. Windows only.

Price: free.

2. Barside

Barside is a free desktop app for Windows, Linux and Mac that keeps notes, tasks, reading and recordings in one library on your computer. Its screen text tool reads any part of the screen and lets you copy the text, or keep the capture in a note. Read the full guide to copying text from the screen with Barside.

Best for: people who want to keep what they capture, such as error messages, slides, receipts and quotes, and find it again later. Also the best pick here for Linux users who want captures saved with their notes.

Strengths:

• Press Alt+Shift+S in any app (or use the toolbar button, the search bar or Capture screen text in the tray). Every monitor freezes, and you drag across the text you want. Enter captures the selection or the whole monitor, the arrow keys move the selection (with Shift, 10 pixels at a time), Ctrl with the arrows resizes it, and Esc cancels.
• The Text from your screen window shows the image and the recognized text, with Copy text, Copy image, Add to current note, Add to a note… and New note.
• A capture you add to a note keeps the image and its text, and text in images inside your notes becomes searchable.
• Hide Barside while capturing (on by default) keeps Barside's own windows out of your captures.
• Text is read on your computer, with Windows OCR on Windows, Tesseract on Linux and Apple's Vision framework in Barside for Mac. Nothing is uploaded, and a capture you close without adding to a note is deleted.

Limits: there is always a result window, so a plain copy takes one more click than PowerToys Text Extractor. It reads text only: no QR codes, table structure or translation. A selection stays on one monitor. On Linux, install Tesseract first (sudo apt install tesseract-ocr), and on Windows, add a language with OCR support if Settings → Screen capture says text recognition is missing (its Language settings button takes you there). Barside for Mac needs no setup: it reads text with Apple's Vision framework. On a Linux desktop that uses Wayland, Alt+Shift+S works while an X11 app is active; to use it everywhere, copy its command in Settings → Hotkeys and aliases and add it as a shortcut in your desktop's keyboard settings.

Price: free.

3. Snipping Tool (Windows 11)

The Snipping Tool that comes with Windows 11 can read the text in a screenshot. Take a snip, open it in the Snipping Tool, and select Text actions. You can then select and copy part of the text, use Copy all text, or use Quick redact to hide email addresses and phone numbers before you share the image. Microsoft says all text recognition runs locally on your device.

Best for: occasional copies when you do not want to install anything.

Strengths: already on your computer, local, and it doubles as a screenshot and redaction tool.

Limits: it takes more steps than a dedicated OCR tool (snip, open, Text actions, copy), and it is Windows 11 only.

Price: free with Windows.

4. NormCap

NormCap is a free, open-source "OCR-powered screenshot tool to capture text instead of images", for Windows, macOS and Linux. You select an area, and the text goes straight to your clipboard.

Best for: people who use more than one operating system, and anyone who needs to read QR codes and barcodes from the screen.

Strengths:

• Tesseract language models for many languages, and you can add more.
• Its Parse Text mode tidies the result: it keeps or joins lines and paragraphs, and turns captured email addresses and URLs into clean lists.
• Works with several monitors and different scaling settings.
• Detection runs on your device, and NormCap says it collects no telemetry.

Limits: it is built to put text on your clipboard, not to keep your captures. Accuracy depends on Tesseract and on the language models you install.

Price: free (GPLv3).

5. Text Grab

Text Grab is a Windows OCR app for Windows 10 and 11, by the same developer whose code PowerToys Text Extractor is based on. It goes further than Text Extractor, with four modes: Full-Screen (select a region), Grab Frame (a movable window you place over content), an Edit Text window for cleaning up results, and Quick Simple Lookup for text you use often.

Best for: power users who capture text all day and want to clean it up, extract patterns or grab tables.

Strengths: a choice of engines (Windows OCR, Tesseract and, in the Store version on Copilot+ PCs, Windows AI), direct text extraction through UI Automation where the app exposes it, patterns and regular expressions for structured data, and tables and lists with formatting through its browser extension. All OCR runs on your computer.

Limits: the app is for Windows only (its free browser extension for tables and lists works in Chromium browsers on any system), and the extra modes take a little learning.

Price: free as a build on GitHub, or US$9.99 in the Microsoft Store, which includes its latest OCR models and local AI (as of September 2026).

6. ShareX

ShareX is a free, open-source screenshot and screen recording tool for Windows, and it includes OCR. It uses the OCR engine included with Windows to read a selected region or an image file, locally. The OCR window lets you correct the text before you use Copy all, and you can open the text in a web service such as Google Translate, Google Search or DeepL if you choose. Process OCR silently skips the window and copies the text straight to the clipboard.

Best for: people who already use ShareX for screenshots and uploads.

Strengths: local recognition, a silent mode that copies without opening a window, and OCR as a step in ShareX's capture workflows.

Limits: Windows only (Windows 10 version 1903 or later for OCR), and ShareX's many settings are a lot to take on just for OCR.

Price: free.

7. Snagit

Snagit, from TechSmith, is a paid screenshot tool for Windows and macOS. Its Grab Text feature reads the text in a selection or an existing image, and its Edit Text feature lets you change the text inside the screenshot itself: its color, font and size, or the words.

Best for: people who make documentation and need to fix or update text in screenshots, not only copy it.

Strengths: Grab Text during a capture or on any image in Snagit, a plain-text option in the results, and Edit Text for changing text in place.

Limits: a paid subscription. TechSmith's support article lists a short set of supported languages for Grab Text, so check it if you read other languages.

Price: US$39 a year for an individual subscription, with a 15-day free trial (as of September 2026).

8. Click to Do (Copilot+ PCs)

Click to Do is built into Windows 11 on Copilot+ PCs. Press the Windows key and click, or press Win+Q, and Windows looks at what is on your screen so you can select text or an image and act on it. For text, the basic actions are Copy, Open with and Search the web, which searches with Microsoft Edge and Bing. When you select 10 or more words of English and are signed in, it can also summarize, make a bulleted list or rewrite the text.

Best for: people who own a Copilot+ PC and want to copy or act on text with nothing to install.

Strengths: it is part of Windows, you can also open it from the Snipping Tool or Windows Search, and Microsoft says the analysis of the screenshot is always performed on your device. Click to Do does not keep any content from your screen after it finishes the action you asked for.

Limits: it needs a Copilot+ PC (or an eligible Cloud PC), which means specific hardware such as a 40 TOPS NPU and 16 GB of memory. Some of its actions, such as visual search with Bing, are not offered in the European Economic Area, and some work differently in China.

Price: free with Windows on a Copilot+ PC.

Privacy: where your screen goes

Screens hold private things: messages, customer records, passwords in plain sight. It is worth knowing where each tool reads the text.

Keeping text off the internet is only half of privacy. The other half is what stays on your computer afterwards. Text Extractor, NormCap and the Snipping Tool's text actions leave the text on your clipboard, and a clipboard history manager may keep it. Barside keeps a capture only when you add it to a note, and its own clipboard history is off until you turn it on.

How to choose a screen OCR tool

Two tools can work well together: PowerToys Text Extractor for quick copies, and Barside for the captures worth keeping, because a screenshot of an error message, a slide or a receipt in a note can be found again by a word inside it, with the search box above your notes or by asking your library a question. The search bar (Alt+Shift+Space) finds the note by its title and text, and typing ocr there, then Enter, starts a new capture.

Tips for accurate results

• Make the text bigger first. Zoom the page or the image before you capture. Larger text is read more reliably, especially small print and dense tables.
• Install the right language. Tools that use Windows OCR need the language's OCR pack. Tesseract-based tools need the language data. Check this first when a tool returns nothing.
• Capture only what you need. A tight box around one column reads better than a whole screen with sidebars, icons and a busy background.
• Proofread numbers and code. OCR can confuse similar characters, such as the letter O and the digit 0, or a lowercase l and the digit 1. Check order numbers, amounts and commands before you use them.
• Use the keyboard for precision. In Barside, the arrow keys move the selection one pixel at a time and Ctrl with the arrows resizes it, which helps with small text.

Other ways to copy text from images

If the text is already in a picture on your computer rather than on the screen, you have more options. Microsoft OneNote on Windows can copy text from a picture you insert: right-click the picture and choose Copy Text from Picture, though Microsoft notes that the command may not appear right away for complex pictures, and some results can take up to 24 to 48 hours. If you are leaving OneNote for a local app, see the OneNote alternative guide.

Text you copy often ends up in a clipboard manager, and some of them keep everything you copy. Compare the best clipboard managers for how each one handles private text. And if what you need is a video rather than a screenshot, see the best screen recorders.

Works well with screen OCR