Skip to content

v0.3.0

Pre-release
Pre-release

Choose a tag to compare

@BrycensRanch BrycensRanch released this 26 Jun 17:38
· 1140 commits to develop since this release
ea6f33f

v0.3.0 (2025-06-26)

You can now upload and download your HistoryItems directly and instantly open their URLs, shortened links, their deletion links, file paths, and folders. We just upgraded to PaddleOCR v5 (PaddleSharp v3), which increases the performance of the OCR model. Additionally, you can now run PaddleOCR on most Linux distributions*, and macOS. To accompany the support, we've created an OCR window for it that has support for English, Chinese (Simplified), Chinese (Traditional), Spanish, Arabic, Korean, Hindi, Japanese, Telugu, Tamil, Russian, Portuguese, Turkish, German*, French (Revolution).

The ugly "sidebar" we had has been ejected from the building. Now, there's an actual sidebar that mimics ShareX's sidebar, somewhat.

Additionally, we've removed the dependency on LibVLC, which didn't make much sense for us. After giving it thought, playback inside SnapX is a waste of everyone's time. On Linux, we have mpv, haruna, and VLC. On Windows, it ships Movies & TV ootb and you have VLC as an option. On macOS, you have the integrated video player with VLC as an option. For sound playback, we now use ffplay from FFMPEG.

We've also made SnapX far more stable. The annoying HistoryItem flashing is gone, and the region selector won't crash the app if screenshotting fails. It's simply a much smoother and more reliable experience.

You'll find smarter build processes, with us no longer putting *.so files in your $PATH to appease .NET. We now use a wrapper script, as we should have in the first place. For Windows users, this release includes better path handling and disables the wrapper script on Windows. Plus, we've fine-tuned performance by optimizing build times on Windows and ensuring a smoother experience when launching features like the About Window. We've also updated various dependencies for improved stability and security.

Additionally, I have added a LogViewer for ease of access.

I also switched SnapX's versioning tool to GitVersion, which was designed to be SemVer compliant. It also doesn't panic when building SnapX without a git repository. Finally, we've improved our documentation, making it clearer for packagers and easier to understand our versioning.

SnapX is now ~825 MiB uncompressed (~411 MB zipped), with the majority of that coming from our integration of PaddleOCR, a powerful offline AI text recognition engine. This allows SnapX to deliver fast, private OCR without needing any cloud services. While the increase is most noticeable on Linux (previous builds were ~75 MiB), it's less significant on Windows and macOS. When I asked in our Discord, the community overwhelmingly understood and accepted the size tradeoff in exchange for better cross-platform support and advanced OCR. If you feel that is too large for your liking, you can set the DisableOCR=1 environment variable when building SnapX, and it will once again be ~75 MiB.

"In open source, we feel strongly that to really do something well, you have to get a lot of people involved." - Linus Torvalds

Known Issues

  • Much higher memory usage than ShareX. This is because even though SnapX is still under active development, it utilizes more of your computer's memory than before, such as caching images in memory to keep the UI responsive.
  • The settings window's buttons do nothing yet, besides the Config folder button.
  • PaddleOCR can't be run on musl-based Linux distributions, Windows 11 ARM64, and FreeBSD.
  • On Linux, PaddleOCR libraries contain an invalid RPATH, but still run. This still causes COPR builds to fail. It will be fixed soon after this release.

Screenshots

image
image
image
image
image


Features

Bug Fixes

Build

Dependencies

Documentation

Misc


Full Changelog: v0.2.1...v0.3.0