NormCap v0.4.4

The prebuild packages are shipped with support for English only. To install additional languages, click the settings icon and in the "Languages" section of the menu "add/remove...".

If you installed NormCap as Python package, refer to the online documentation on how to install additional language for Tesseract on your system.

The quality of detections is mainly determined by Tesseract, an open source project for OCR mainly of printed text. NormCap builds upon this great library and can influence the quality only a bit by preprocessing the images regarding screen specific properties. You, the user, have much more influence on the quality by paying attention to certain aspects:

1. Select the correct language – To specify the correct language(s) in the settings menu is the very important. But keep in mind: selecting multiple languages at once slows down the recognition a bit.
2. Avoid selecting decorations – The text detection is very sensible to elements other than text, for example borders, lines or icons. If possible, do not select those:
   
   
3. Avoid heterogeneous colors – If the text you want to recognize has different backgrounds, it can help to select those portions of text separately:
   
4. Zoom in – If possible, enlarge the region you want to detect, before selecting it with NormCap. This has a huge effect in situations, where zooming in doesn't decrease the quality, e.g. zooming into webpages, PDF documents or images which are not yet at their full resolution. It has little effect, if enlarging decreases the quality, e.g. for images with low resolutions.
5. Select more text – Sometimes Tesseract struggles with recognizing text with only very few characters, like a single word. In this case, selecting a larger portion of text can improve the accuracy.
6. Try different language models – The prebuild NormCap packages are using tessdata-fast models, which offer a very good accuracy to speed compromise. But you can also try the slower and larger models from tessdata or tessdata-best instead.
   To find the directory in which you have to put the manually downloaded models navigate to the "Language" section of NormCap's settings, then click "add/remove" and finally "View tessdata folder in File Manager".
7. Report examples – If the results are still bad, please submit a screenshot of the text you are trying to recognize as an issue. Chances are small, but we might be able to improve something.

No. The text recognition is performed offline using the OCR framework Tesseract.

No telemetry data is collected and by default also no other data is sent anywhere.

However, some (optional) features require accessing resources on the web and therefore expose minimal information (like IP address) to the accessed server:

• If you enable the check for updates on start, the application checks NormCap's releases page on GitHub.
• If you download additional languages, they are being fetched from GitHub, too.

Yes, but you'll have to configure it on your own using your operating system's tools:

• Linux:
  • Ubuntu/Gnome: Set Keyboard Shortcuts
  • Manjaro/Xfce, Plasma: Keyboard Shortcuts
• macOS: Set a keyboard shortcut to open Mac apps
• Windows 10: Open programs with keyboard shortcuts in Windows 10

Ask your preferred search engine for support or different ways to do this.

To launch NormCap minimized to system tray without triggering a capture, use the `--background-mode` command line flag. This is e.g. useful, if you want to autostart NormCap after system start.

Note that `--background-mode` automatically enables the setting "Keep in system tray"!

NormCap's MSI installer is much larger than the ones for Linux and macOS because the included Tesseract binaries are larger.

NormCap used to include a set of smaller binaries, which resulted in more consistent installer file sizes for all operating systems. Unfortunately, they seemed to lack some dependencies and lead to issues on some Windows systems. Hopefully, the larger binaries will fix this issue. This decision is not set in stone, so please feel free to leave some feedback regarding this topic!

Some applications offer similar features like NormCap and might be a good or even better alternative for you: It completely depends on your use case and requirements. Some similar open-source Projects are:

• TextSnatcher
• GreenShot
• TextShot
• gImageReader
• Capture2Text
• Frog
• Textinator
• Text-Grab
• dpScreenOCR

You can start NormCap from the command line with the verbosity option to get an idea about what it is doing. The command to run depends on the OS and install method.

• With Python package:
  

  Run: normcap -v debug

• On Windows (MSI):
  

  Run: %LOCALAPPDATA%\Programs\dynobo\NormCap\NormCap.exe -v debug

• On macOS (DMG):

  Run: /Applications/NormCap.app/Contents/MacOS/NormCap -v debug

• On Linux (AppImage):

  Run: ./NormCap-{version}-x86_64.AppImage -v debug

• On Linux (FlatPak):

  Run: flatpak run --command=normcap com.github.dynobo.normcap -v debug

• On Linux (AUR):

  Run: normcap -v debug

Please try to take a look at the debug information. It might provide enough information for you to solve the issue for yourself. If it doesn't help you, don't hesitate to report your problem description together with debug information.

Is your display environment Gnome Shell? Then you probably need to install a Gnome Shell extension to support showing applications in the top bar, e.g.:

• AppIndicator Support
• Ubuntu AppIndicators
• Tray Icons
• Tray Icons: Reloaded

NormCap's notifications depend on the system tray functionality. If you start NormCap, but its Icon doesn't appear in the system tray, proceed like in the question above.

In case you get an output like this...

$ normcap
QtFatalMsg - This application failed to start because no Qt platform plugin
could be initialized. Reinstalling the application may fix this problem.
Available platform plugins are: eglfs, linuxfb, minimal, minimalegl, offscreen,
vnc, wayland-egl, wayland, wayland-xcomposite-egl, wayland-xcomposite-glx, webgl, xcb.

... the chances are, this can be solved by installing additional dependencies:

• Arch (Wayland): pacman -S qt6-wayland
• Arch (Xorg): pacman -S libxcb xcb-util-cursor
• Debian/Ubuntu (Wayland): apt install qt6-wayland
• Debian/Ubuntu (Xorg): apt install libxcb1 libxcb-cursor
• Fedora (Wayland): dnf install qt6-qtwayland
• Fedora (Xorg): dnf install libxcb xcb-util-cursor

This is behavior was observed only on KDE + Wayland (#422) so far, and only when NormCap was started via a keyboard shortcut / key binding. The root cause is still unknown, if you have any information or ideas, please comment in the ticket above.

Strangely, a workaround seems to be to configure the keyboard shortcut via "System Settings" → "Shortcuts" → "Add command" and configure the command in a way, that it pipes the output to somewhere, e.g. /dev/null:

normcap 2>&1 | tee /dev/null

This is not a NormCap issue but a requirement for AppImages. You need to make sure, that the FUSE library is installed on your system. E.g. on Ubuntu 22.04 you need to run sudo apt install libfuse2. See this blog post for details.

The application used to package the AppImage has a number of runtime requirements. One of those requirements is libcrypt.so.1, which should be provided by most modern Linux distributions, as it is mandated as part of the Linux Standard Base Core Specification. However, some distributions don't include libcrypt.so.1 as part of the base OS configuration. This can usually be fixed by installing the libxcrypt-compat package.

This issue usually occurs on the after installing NormCap, either for the first time or after an update.

This is a known issue related to macOS's permissions settings: If NormCap doesn't have the system's permission to take a screenshot, an empty desktop will be shown.
(To be precise: NormCap doesn't know that it lacks permissions, tries to take a screenshot nevertheless, which results in a screenshot of the empty desktop.)

Steps to solve this:

1. Close NormCap, if it is running.
2. Navigate to "System Preferences" → "Security & Privacy" → "Privacy" → "Screen Recording" → "Click unlock".
3. Do you already see "NormCap" on the right side? If yes, "remove" [ – ] it. Un-ticking the checkbox is not enough!
4. Click "add" [＋] → "Applications" → "NormCap".
5. Confirm that you see NormCap on the right side with a checkmark in front of it.
6. Start NormCap, it should work now.
7. You might need to repeat those steps after installing a new version of NormCap.

Hopefully, this cumbersome user experience can be improved in a future release.

To verify, if this is the issue you are facing, run the docker image interactively and try to run the linuxdeploy-*.AppImage file there:

$ docker run -it \
  --volume /home/<USER>/<PROJECT PATH>/normcap/linux:/app:z \
  --volume /home/<USER>/.briefcase:/home/brutus/.briefcase:z \
  --env VERSION=0.2.0 briefcase/eu.dynobo.normcap:py3.9 \
  /bin/bash
$ /home/brutus/.briefcase/tools/linuxdeploy-x86_64.AppImage

If that results in a `No such file or directory` error, according to this issue and this one a workaround is to correct the "magic" bytes of the AppImage. This worked for me:

sed '0,/AI\x02/{s|AI\x02|\x00\x00\x00|}' -i linuxdeploy-x86_64.AppImage

1. Install dependencies:
   Note: wl-clipboard can be omitted on X11 based systems
   
   
   • Ubuntu/Debian:

     sudo apt install build-essential tesseract-ocr tesseract-ocr-eng
                         libtesseract-dev libleptonica-dev wl-clipboard

   • Arch:

     sudo pacman -S tesseract tesseract-data-eng wl-clipboard

   • Fedora:

     sudo dnf install tesseract wl-clipboard

   • openSUSE:

     sudo zypper install python3-devel tesseract-ocr
                         tesseract-ocr-devel wl-clipboard

2. Install NormCap:

   pip install normcap

3. Run:

   ./normcap

1. Install dependencies:

   brew install tesseract tesseract-lang

2. Install NormCap:

   pip install normcap

3. Run:

   ./normcap