OCR Text Extractor for GNOME: Screenshot to Text

A GNOME Shell extension for OCR on your screen: select an area, and the text in it (screenshot to text) is copied to the clipboard. It is the GNOME equivalent of PowerToys Text Extractor on Windows, and works with any language Tesseract supports.

1. Start	2. Select	3. Paste
Press Super+Shift+T, or click the panel icon and choose Extract Text from Screen	Drag over the text, the capture happens when you let go	The text is on your clipboard

Features

• Uses GNOME's own screenshot overlay, reduced to area selection: the screen freezes, you drag, you are done
• Works with any language Tesseract has installed, and with several at once (eng+tam)
• Bring your own models: point the extension at a folder of .traineddata files
• Quiet: a short on-screen message on success, a notification only when something needs fixing
• Nothing is left behind: the temporary screenshot is deleted after reading

Requirements

• GNOME Shell 45 to 50
• Tesseract and the language data you need

The extension does not install anything itself. When Tesseract or a language is missing, it tells you the exact command for your distribution, in the notification and in Preferences → Text Recognition (with a copy button).

Distribution	Tesseract	A language, e.g. Tamil (tam)
Ubuntu, Debian	sudo apt install tesseract-ocr	sudo apt install tesseract-ocr-tam
Fedora	sudo dnf install tesseract	sudo dnf install tesseract-langpack-tam
Arch	sudo pacman -S tesseract	sudo pacman -S tesseract-data-tam
openSUSE	sudo zypper install tesseract-ocr	sudo zypper install tesseract-ocr-traineddata-tam

Adding a language

English works out of the box. To read another language:

1. Find its Tesseract code, for example deu (German), tam (Tamil), sin (Sinhala), chi_sim (Chinese, simplified). See the full list.
2. Install the language pack with your package manager, using the table above. For Sinhala on Ubuntu: sudo apt install tesseract-ocr-sin.
3. Open Preferences → Text Recognition, press the refresh button, and tick the language. Tick several to read mixed-language text.

The extension lists whatever Tesseract has installed, so no extension update is needed for new languages.

A language your distribution does not package

1. Download its .traineddata file from tessdata_best (best quality) or tessdata_fast (faster).
2. Put it in a folder of your own, for example ~/tessdata, together with eng.traineddata and any other language you still want. Tesseract reads one folder only.
3. In Preferences, enter that folder in Tessdata folder and press apply, then tick the language.

Installation

1. Install Tesseract first, as described in Requirements. The extension does not install it for you.

2. Install the extension, either way:

   • From extensions.gnome.org: open the page and switch it on.

   • From source:

     git clone https://github.com/imshaaz21/text-extractor-gnome.git
     cd text-extractor-gnome
     make install

     Log out and back in, then enable it: gnome-extensions enable text-extractor@imshaaz21.github.com

Uninstallation

1. Remove the extension. Use the Extensions app (Remove), or run gnome-extensions uninstall text-extractor@imshaaz21.github.com. If you installed from source you can also run make uninstall.

2. Optional: reset its settings:

   dconf reset -f /org/gnome/shell/extensions/text-extractor/

3. Optional: remove Tesseract if nothing else uses it. The extension never installed it, so it stays until you remove it with your package manager, for example sudo apt remove tesseract-ocr on Ubuntu or sudo dnf remove tesseract on Fedora. Language packs you added are separate packages.

Usage

1. Press Super+Shift+T, or click the panel icon and choose Extract Text from Screen.
2. Drag over the text.
3. Paste it anywhere.

Press Esc to cancel.

Preferences

Setting	What it does
Languages	Tick one or more installed languages. Mixed-language text works when several are ticked.
Tessdata folder	Use your own .traineddata files, for example from tessdata_best. Tesseract reads one folder only, so this replaces the system folder: put every language you need in it.
Extract Text Shortcut	Keyboard shortcut that starts a capture (default Super+Shift+T). Click the row and press new keys to change it, or use the clear button to disable it. Works with the panel icon hidden.
Show Panel Indicator	Hide or show the panel icon.

FAQ

Is there a PowerToys Text Extractor for Linux? This extension does the same job on GNOME: you select an area of the screen and the text in it is copied to the clipboard.

How do I copy text from an image or from the screen on Ubuntu, Fedora or Arch? Install Tesseract, install this extension, press Super+Shift+T (or click its panel icon) and drag over the text. The text is on your clipboard.

Does it work on Wayland? Yes. It uses GNOME's own screenshot overlay, so it does not depend on X11 tools.

Which languages does it read? Any language Tesseract has data for, including several at once. See Adding a language.

Does it send my screen anywhere? No. Recognition runs locally with Tesseract, and the temporary screenshot is deleted afterwards.

Troubleshooting

"Missing: ..." message. Run the command it shows, then press the refresh button in Preferences.

The text is wrong or empty. Select a larger area and make sure the right language is ticked. Tesseract reads clean, high-contrast text best.

Something else. Open an issue with the output of journalctl -f -o cat /usr/bin/gnome-shell while you reproduce it.

Development

make nested   # run the code in extension/ in a throw-away GNOME Shell
make pack     # build dist/*.shell-extension.zip

extension/
├── extension.js   panel menu, screenshot to text to clipboard
├── prefs.js       preferences window
├── ocr.js         Tesseract helpers shared by both
├── metadata.json
└── schemas/       GSettings schema

See CONTRIBUTING.md to get involved, docs/README.md for debugging tips and CHANGELOG.md for release notes.

License

GNU General Public License v3.0