Barside
Download
Barside/Content/Capture · Screen text

Capture · Screen text

Copy text from the screen with OCR, from any app, image or video

Press Alt+Shift+S, drag across any part of any screen, and Barside reads the text with the OCR built into your computer. Copy the text or the image, or keep the capture in a note, where its words become searchable. Nothing is uploaded.

Feature guideUpdated 9 min read

Some text cannot be selected: an error message in a dialog, the words in a screenshot, a paused video, a slide someone is sharing, a scanned page. OCR (optical character recognition) reads that text from the picture. With Barside you can copy text from the screen in any app: press Alt+Shift+S, drag across the text, and Barside shows what it read, ready to copy. The OCR runs on your computer, so the capture never leaves it. It uses the text recognition built into Windows or Tesseract on Linux, and Barside for Mac uses Apple's Vision framework.

You can also keep a capture in a note, and every screenshot in your notes becomes searchable by the words in it.

How to copy text from the screen (OCR) with Barside

  1. Press Alt+Shift+S in any app. You can also choose Capture screen text in a note's toolbar, in the search bar (type ocr) or in the tray menu.
  2. Every monitor freezes on a snapshot, and the screen dims. Drag a rectangle over the text you want. The size of the selection shows in pixels as you drag.
  3. Let go. Barside crops the snapshot at the monitor's full resolution and reads its text.
  4. The Text from your screen window opens, with the capture, the Recognized text and its number of lines.
  5. Choose Copy text and paste it wherever you need it.

Changed your mind? Press Esc, right-click, or press the shortcut again to cancel. If the capture has no text, the window says No text found, and you can still keep the image.

What you can do with a capture

  • Copy text puts the recognized text on the clipboard.
  • Copy image copies the picture instead, for a chat or a document.
  • Add to current note adds the image to the note you have open.
  • Add to a note… opens a search of your notes, most recently changed first. Pick one with the arrow keys and Enter.
  • New note saves the capture in a new note titled Screen capture, with the date and time.

After adding, Barside opens the note. The capture sits there as an ordinary image, with the text Barside read stored alongside it, so a search can find it later. Captures you close without adding to a note are deleted, so they never pile up in your media folder.

Copy text from anything · Open the video

Keyboard and mouse controls

ActionHow
Select an areaDrag, then let go
Start a selection without the mouseAny arrow key (a box appears in the middle)
Capture the selection, or the whole screenEnter
Move the selectionArrow keys (with Shift, 10 pixels at a time)
Resize the selectionCtrl and the arrow keys
CancelEsc, right-click, or the shortcut again

An on-screen hint reminds you: drag across the text you want, Enter captures the whole screen, Esc cancels. A selection stays on one monitor. If your monitors use different scaling, Barside maps the selection onto each monitor's real pixels, so small text is read at full resolution.

Screenshots in your notes become searchable

Screen text is not only for captures. When you open a note with images in it (PNG, JPG, BMP, GIF or WebP), Barside reads their text in the background with the same OCR and stores it with each image. From then on:

  • The search above your notes finds the note by the words inside its screenshots.
  • Library search matches them too, so AI search can use them as sources when you ask your library a question.

A screenshot of a receipt, a photo of a whiteboard or a slide pasted into a note can be found later by what it says. On Windows, Win+Shift+S copies a screenshot and Ctrl+V pastes it into the note.

This is different from Read screenshots with AI in Library → Sources, which sends images to the AI you connected so it can describe them. Screen text never does that: it only reads the letters, on your computer. The assistant's screen understanding is different too: when you ask it to look, it sends a picture of a window or monitor to your AI.

Copy text from anything on your screen · Open the video

Set up text recognition

Barside uses the text recognition your system already has:

  • Windows 10 and 11: Windows OCR, in the languages installed in Windows. If none of them includes text recognition, Settings → Screen capture says so, and Language settings opens Windows' language page, where you add a language that includes optical character recognition.
  • Linux: Tesseract, the open-source OCR engine. The .deb recommends it. With the AppImage, install it yourself, for example with sudo apt install tesseract-ocr on Ubuntu or Debian, then reopen the settings page. Barside reads with the Tesseract languages you have installed, up to three at a time. Add a language with its package, such as tesseract-ocr-fra for French.
  • Mac: Barside for Mac uses Apple's Vision framework, so there is nothing to install. It needs macOS 12 or later.

Settings → Screen capture also lists the languages it can read, has a Capture button to try it, and has Hide Barside while capturing, on by default, which leaves Barside's own windows out of your captures. On Windows they stay visible on your screen while they are left out; on Linux they hide for the moment of the capture, and in Barside for Mac, they turn transparent for that moment.

To change Alt+Shift+S, open Settings → Hotkeys and aliases. On a Wayland desktop, global hotkeys work while an X11 app is active. To use one everywhere, copy its command there and add it as a shortcut in your desktop's keyboard settings. ## Ways to use screen text

  • Error messages in dialogs that do not let you select them. Capture, copy, then search the web or paste into a bug report. A screen recording of the steps can sit in the same note.
  • Paused videos and webinars. Grab a command, a web address or a slide's bullet points.
  • Scanned PDFs and photos, where the text is only pixels.
  • Invoices and receipts. Copy an invoice number or a total into a spreadsheet.
  • Code in screenshots that someone sent you in a chat.
  • Meetings. Capture a slide someone shares and add it to your meeting note, where its words become searchable.

Text you copy this way also lands in your clipboard history when that is on, so you can paste it again later.

Tips for clean results

  • Select only the text you need. A tight rectangle around one block usually reads better than a whole busy screen.
  • Zoom in first when the text is tiny, then capture.
  • Proofread numbers and names. OCR can confuse similar characters, such as 0 and O, or 1 and l. Check amounts before you pay anything.
  • Keep the image with New note when you may need the original. The text alone loses the layout.

Troubleshooting

  • Nothing happens when you press Alt+Shift+S. Another app may hold that shortcut. Settings → Screen capture tells you when it does; choose another hotkey in Settings → Hotkeys and aliases, or start from a note's toolbar or the tray menu. On Wayland, remember that global hotkeys work while an X11 app is active.
  • "Windows text recognition is not installed" or "Text recognition needs Tesseract". Add a Windows language with text recognition or install Tesseract on Linux, as described above, while Barside for Mac needs no setup, only macOS 12 or later.
  • "Text recognition took too long." Recognition stops after a minute. Select a smaller area and try again.
  • Barside's own window shows up in the capture. Turn on Hide Barside while capturing in Settings → Screen capture.
  • The screen is locked. Captures do not run while the screen is locked. Unlock it, then try again.
  • A capture failed. The window says so and offers Try again. If it keeps failing, the Send to support button next to the error sends Barside a report of the error. Reports never include your notes, your clipboard or your files.

Privacy

Alt+Shift+S captures only the part of the screen you select. Your computer's own text recognition reads it, and nothing is uploaded. A capture you add to a note becomes part of that note, and of its backups; a capture you close is deleted. Capture preferences stay on this computer, outside your backups. Read how screen text works in the privacy guide.

How it compares

Windows has other ways to copy text from the screen. The Snipping Tool's Text actions read the text in a screenshot on your device, with Copy all text and a quick redaction of email addresses and phone numbers. PowerToys Text Extractor copies the text in a selected area straight to the clipboard, using the Windows OCR language packs.

ToolSystemsShortcutWhat you get
BarsideWindows, Linux and MacAlt+Shift+S, changeableText or image, copied or kept in a note
Snipping ToolWindowsWin+Shift+S, then Text actionsSelected or all text, with quick redaction
PowerToys Text ExtractorWindowsWin+Shift+T by defaultText copied to the clipboard

Barside and the Snipping Tool read the text on your device, and Text Extractor uses the OCR language packs installed in Windows. Barside adds Linux and Mac, keeps the capture in a note when you want it, and makes the screenshots across your notes searchable. For more options, see the best OCR tools to copy text from the screen.

Works well with

FAQ

Questions and answers

How do I copy text from a screenshot or an image on my screen?

Press Alt+Shift+S, drag a rectangle over the text, and choose Copy text in the window that opens. It works on anything you can see: images, paused videos, PDFs, apps and dialogs. The text recognition runs on your computer.

Does Barside upload my screen to read the text?

No. Barside reads the capture on your computer: Windows uses its built-in text recognition, Linux uses Tesseract, and Barside for Mac uses Apple's Vision framework. Nothing is uploaded, and a capture you do not add to a note is deleted.

Which languages can Barside's OCR read?

The languages your system can recognize. On Windows, those are the languages installed in Windows that include text recognition. On Linux, Barside reads with the Tesseract languages you have installed, up to three at a time; add one with its package, such as tesseract-ocr-fra for French. In Barside for Mac, they are the languages Apple's Vision framework supports. Settings → Screen capture lists the languages Barside found.

Why does Barside say text recognition is not installed?

On Windows, none of your installed languages includes text recognition: open Windows Settings → Time & language → Language & region and add one that does. On Linux, install Tesseract, for example with sudo apt install tesseract-ocr, then reopen Settings → Screen capture. Barside for Mac reads text with no setup, as long as the Mac runs macOS 12 or later.

Can I change the Alt+Shift+S shortcut?

Yes, in Settings → Hotkeys and aliases. You can also start a capture from a note's toolbar, from the search bar (type ocr) or from the tray menu. On a Wayland desktop, add the command copied from Hotkeys and aliases as a shortcut in your desktop's keyboard settings to use it everywhere.

Does screen text work with more than one monitor?

Yes. Every monitor freezes when you start a capture, so you can select text on any of them. A selection stays on one monitor, and Barside reads it at that monitor's full resolution, even when your screens use different scaling.

Can I search the text inside screenshots in my notes?

Yes. When you open a note, Barside reads the text in its images (PNG, JPG, BMP, GIF or WebP) in the background and stores it with each image. The search above your notes and Library search then find the note by the words in its screenshots.

Is screen text available on a Mac?

Yes: Barside for Mac is out. Screen text also works on Windows 10 and 11 and on 64-bit Linux. On a Mac, it reads text with Apple's Vision framework, with nothing to install.

Sources

Where the facts come from

Checked on September 30, 2026. Products change: if something here is out of date, tell us at support@barside.space.

  1. Use Snipping Tool to capture screenshots (Microsoft Support)
  2. PowerToys Text Extractor (Microsoft Learn)
  3. Tesseract OCR