Uahh/ToastFish

ToastFish is a Windows desktop application designed for vocabulary memorization and study tracking. It provides a user interface for managing flashcards in English, Japanese, and custom languages, utilizing an SM2+ algorithm to schedule reviews based on user feedback. The system supports audio playback for word pronunciation, tracks learning progress, and allows for the import and export of word lists. Additionally, it offers background operation via system tray integration and global hotkey support.

Users can now recite English, Japanese, and custom imported words via push notifications with interactive buttons (remember/skip/voice). The system tracks learning progress using an SM2+ algorithm, updating word status and review intervals based on user feedback. Additionally, users can export their learning logs to Excel files for each language type, and import custom word lists from Excel spreadsheets.

Model/PushControl · high confidence

Introduced new classes for handling audio playback and downloading. The \MUSIC\ class in \PlayMp3.cs\ provides a wrapper around Windows MCI API to play, pause, and stop audio files, while \DownloadMp3.cs\ implements HTTP-based MP3 downloading and caching logic. This adds the ability to fetch and play MP3 audio files directly within the application.

Model/Mp3 · high confidence

The project now includes a Windows application manifest (app.manifest) that configures the application's execution level as 'asInvoker' and specifies compatibility with Windows 10. Additionally, assembly metadata (AssemblyInfo.cs) and resource settings files have been added to the Properties folder, establishing the application's identity, versioning, and localization support.

Properties · high confidence

The application now supports global hotkeys and automatic startup. A new \HotKey\ class enables registering and handling keyboard shortcuts via Windows API, while a \StartWithWindows\ utility allows the application to create a shortcut in the Windows startup folder, enabling the app to launch automatically when the user logs in.

Model/StartWithWindows · high confidence

The application now includes new localized resources for icons and media files. Specifically, the \IconChika\ resource file and its designer code have been added to the \Resources\ directory. This introduces four new embedded resources: \chika16\ and \chika64\ (16x16 and 64x64 pixel icons), \inami\ (a byte array resource, likely a database or data file), and \mute\ (a byte array resource, likely an audio file). These resources are now accessible via the strongly-typed \IconChika\ class for localization and UI rendering.

Resources · high confidence

The project is initialized with the core application structure, including the Visual Studio solution and project files, .NET 4.7.2 configuration, and the MIT license. The README provides user-facing documentation for the word-memorization tool, outlining features such as vocabulary selection, study tracking, and custom content import.

(repo-wide) · high confidence

Added a new MVVM structure to the application by introducing the ToastFishModel class, which serves as the primary view model with a Push command, and a ViewModelLocator to register and resolve it via the CommonServiceLocator. This establishes the data-binding foundation for the UI layer.

ViewModel · high confidence

The application now includes a main window (ToastFish.xaml) that displays version information, support links, and a notification area. The corresponding code-behind (ToastFish.xaml.cs) implements the primary user interface logic, including a system tray icon for background operation, a right-click context menu for settings and imports, and global hotkey listeners (Alt+D1-D4, Alt+Q, Alt+\~) to trigger actions like starting or pushing words. This establishes the core interaction model for the desktop application.

View · high confidence

The ToastFish application now uses a standard .NET 4.7.2 C\# project file (\.csproj\) to define the build configuration, target framework, and source file inclusions. This establishes the foundational build system for the application, enabling compilation and packaging of the Windows desktop app.

(dependencies) · medium confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 38 → 38 (+0.0)
• Rubric changed (rubric-2026.08.18 → rubric-2026.08.19) — scores are not directly comparable.

• Code Health 65 → 65 (+0.0)
• Architecture 98 → 98 (+0.0)
• Maturity 39 → 39 (+0.0)
• Readiness 36 → 36 (+0.0)
• Security 28 → 28 (+0.0)

• Inconsistent capitalization in method names: 'Puase' (sic) vs 'play'. Additionally, 'Puase' appears to be a typo for 'Pause'.
• Inconsistent naming for hotkey modifier types: 'KeyModifiers' (plural) vs 'KeyModifier' (singular).
• Inconsistent naming for word entity types. Some use 'Word' with a specific qualifier (Jp, Goin, Customize), while others use the base 'Word' type. This suggests a potential lack of a unified 'Word' base class or interface, or inconsistent naming for different word types.
• LLM evaluation failed

• Inconsistent capitalization in method names: 'Puase' (sic) vs 'play'. While 'Puase' appears to be a typo for 'Pause', the casing is inconsistent with 'play' and other PascalCase methods in the codebase.
• Inconsistent naming for hotkey-related classes: 'HotKey' vs 'KeyModifier'. These are related concepts but named differently.
• Inconsistent naming for word types: 'Word', 'JpWord', and 'CustomizeWord' are used. While they might represent different data sources, the naming convention is not uniform (e.g., 'JpWord' vs 'Word').
• The class 'Select' is a poor name for a repository/service class. It is also inconsistent with 'PushControl', 'Mp3', etc.
• The download/install instructions give only one link (a Baidu Pan link) plus a manual install step, with no mention of how to run the app after extraction. (README.md)

• Unchanged — 1 containers · 0 contexts · 0 edges

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Uahh/ToastFish was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.