Xirea

Offline AI Chat Assistant for Android

Features • Screenshots • Installation • Building • Tech Stack • License

---

Overview

Xirea is a fully offline AI chat assistant that runs lightweight language models directly on your Android device. No internet required, no API keys, no data leaving your phone — your conversations stay completely private.

Powered by llama.cpp for efficient on-device inference with GGUF models.

---

Features

• 100% Offline — All AI processing happens on-device
• Fast Inference — Optimized for mobile with dynamic RAM scaling
• Chat History — Persistent local storage with Room database
• Model Management — Download, switch, and delete AI models
• Dark Mode — Beautiful Material3 light and dark themes
• Modern UI — Built with Jetpack Compose
• Privacy First — No data collection, no servers, no tracking

---

Screenshots

Home	Chat	Models	Settings

---

Installation

Requirements

• Android 8.0+ (API 26)
• ARM64 device (arm64-v8a)
• At least 4GB RAM recommended
• Storage space for AI models (500MB - 4GB per model)

Download

Or browse all versions on the Releases page.

---

Building

Prerequisites

• Android Studio Hedgehog or newer
• Android NDK 29.0.14206865
• CMake 3.22.1
• JDK 17

Steps

1. Clone the repository

   git clone https://github.com/Danyalkhattak/Xirea.git
   cd xirea

2. Open in Android Studio

   • Open the project folder in Android Studio
   • Wait for Gradle sync to complete

3. Build Debug APK

   ./gradlew assembleDebug

4. Build Release APK (requires signing keystore)

   ./gradlew assembleRelease

Set signing properties in local.properties:

RELEASE_STORE_FILE=path/to/keystore.jks
RELEASE_STORE_PASSWORD=your_store_password
RELEASE_KEY_ALIAS=your_key_alias
RELEASE_KEY_PASSWORD=your_key_password

The APK will be generated at app/build/outputs/apk/

---

Tech Stack

---

Supported Models

Xirea works with GGUF format models. Recommended models for mobile:

Model	Size	RAM Required
Qwen2.5 0.5B Q4	~400MB	4GB
Qwen2.5 1.5B Q4	~1GB	6GB
Llama 3.2 1B Q4	~700MB	4GB
Phi-3 Mini Q4	~2GB	8GB
Gemma 2B Q4	~1.5GB	6GB

---

Performance Optimization

Xirea automatically optimizes for your device:

Device RAM	Context Size	Batch Size
4GB	512	128
6GB	768	256
8GB	1024	256
12GB+	2048	512

• CPU-only inference for maximum compatibility
• Memory-mapped model loading for reduced RAM usage
• Pre-allocated batch buffers for zero-allocation generation
• Near-greedy sampling for faster token generation

---

Project Structure

xirea/
├── app/
│   ├── src/main/
│   │   ├── java/com/dannyk/xirea/
│   │   │   ├── ai/          # AI engine & llama.cpp wrapper
│   │   │   ├── data/        # Room database & repositories
│   │   │   ├── service/     # Download service
│   │   │   └── ui/          # Compose UI screens
│   │   ├── cpp/             # Native C++ code
│   │   │   ├── llama.cpp/   # llama.cpp library
│   │   │   └── llama_jni.cpp # JNI bridge
│   │   └── res/             # Resources
│   └── build.gradle.kts
├── gradle/
│   └── libs.versions.toml   # Version catalog
└── build.gradle.kts

---

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Author

Danyal Khattak

---

Acknowledgments

• llama.cpp — Excellent C++ inference engine
• Jetpack Compose — Modern Android UI toolkit
• Material3 — Design system

---

Made with by Danyal Khattak