PICAZHU Windows 0.1.1 Alpha
Pre-release
Pre-release
PICAZHU for Windows 0.1.1 Alpha
Alpha maintenance release focused on Settings readability and release-pipeline hardening.
What Changed
- Rebuilt the Settings popup into a full-width scrollable settings surface instead of fixed-width cards.
- Improved readability for labels, provider diagnostics, LM Studio model IDs, and AI indexing status.
- Added a clear selected-model summary under the LM Studio model dropdown.
- Added visible settings toggles for AI and subfolder defaults.
- Expanded footer actions so
Test ConnectionsandSave Settingsdo not clip. - Clarified that
Closediscards unsaved edits andSave Settingsapplies immediately. - Hardened the GitHub staging script so it preserves the existing
.gitfolder when refreshing source.
Downloads
Recommended install:
PICAZHU-Windows-Setup-0.1.1-alpha.exe
Portable/testing option:
PICAZHU-Windows-portable-0.1.1-alpha.zip
Verification:
SHA256SUMS.txt
Security And Trust
The 0.1.1-alpha installer is unsigned. Windows may show an Unknown publisher or SmartScreen warning until PICAZHU is signed with an OV/EV code-signing certificate.
For this release:
- Verify downloads against
SHA256SUMS.txt. - Scan the installer or portable zip with your antivirus before installing if your environment requires it.
- The published installer and portable zip were scanned locally with Microsoft Defender on May 9, 2026, and Defender reported no threats.
Antivirus scans are a useful distribution check, but they are not a substitute for code signing.
Verification Performed
dotnet build .\Picazhu.sln -c Release -m:1dotnet test .\Picazhu.sln -c Release --no-build -m:1dotnet publish .\app\Picazhu.App\Picazhu.App.csproj -c Release -r win-x64 --self-contained false -m:1 -o .\publish\Picazhu.App-win-x64 /p:DebugType=None /p:DebugSymbols=falseBuild-WindowsRelease.ps1- Published executable launched and reported
Responding = True. - Microsoft Defender scan found no threats in the installer or portable zip.
Known Alpha Notes
- Installer is currently unsigned.
- This is still a pre-release for controlled testing.
- Embeddings and hybrid semantic ranking are not complete yet.
- OpenAI/Ollama inference paths are not at feature parity with LM Studio.
- Very large iPhone libraries, locked/untrusted phones, iCloud placeholders, and unplug-during-import scenarios need more real-device QA.