Askimo Desktop

A Kotlin Compose Desktop application for the Askimo project.

Features

• Modern Material Design 3 UI
• Simple chat interface with message bubbles
• User and AI message differentiation
• Responsive layout with scrollable message area

Requirements

• JDK 21 or higher
• Gradle 8.x or higher

Running the Application

To run the application in development mode:

./gradlew desktop:run

Building the Application

To build the application:

./gradlew desktop:build

Localization

Detecting Unused Localization Keys

To find unused localization keys in your properties files:

./gradlew :desktop:detectUnusedLocalizations

This task will:

• Scan all localization keys in messages.properties
• Check usage across both desktop and shared modules
• Generate a detailed report at build/reports/unused-localizations.txt
• Display a summary with any unused keys

Creating Native Distributions

You can create native distributions for different platforms:

# Create DMG for macOS
./gradlew desktop:packageDmg

# Create MSI for Windows
./gradlew desktop:packageMsi

# Create DEB for Linux
./gradlew desktop:packageDeb

Project Structure

desktop/
├── build.gradle.kts          # Build configuration
├── src/
│   └── main/
│       ├── kotlin/
│       │   └── io/askimo/desktop/
│       │       └── Main.kt    # Main application code
│       └── resources/         # Resources (icons, etc.)
└── README.md                  # This file

Architecture

The application uses:

• Compose Multiplatform: For building the UI
• Material 3: For modern design components
• Kotlin Coroutines: For asynchronous operations

Voice Playback (MP3 support)

The optional voice feature (Settings > Voice) plays back synthesized speech via javax.sound.sampled, which only decodes WAV/AIFF/AU natively. OpenAI's TTS API returns MP3-encoded audio, so MP3 playback requires a third-party javax.sound.sampled SPI provider on the runtime classpath:

• Dependency: com.googlecode.soundlibs:mp3spi (declared once in desktop-shared/build.gradle.kts via the mp3spi version-catalog entry). Its POM transitively pulls in tritonus-share + jlayer, so no other libraries need to be added by hand.
• Packaging: no extra jpackage/module-path configuration is needed — it's a plain jar on the classpath (registered via META-INF/services/javax.sound.sampled.spi.*), so packageDmg / packageMsi / packageDeb all bundle it automatically like any other dependency.
• Local (Piper) TTS returns WAV instead, which needs no extra SPI.
• Covered by io.askimo.ui.voice.Mp3DecodingTest in desktop-shared's test suite, which fails loudly with an actionable message if this dependency is ever accidentally removed.