Harry Cartwright - Portfolio

Case Study - Layla Android

Android version of Layla - Private AI chatbot that operates directly on your device without requiring an internet connection.

Industry
Mobile App
Year
Service
Mobile App (Android)

About

Layla Android is the Android version of the groundbreaking personal AI application. It provides the same powerful offline AI capabilities as the iOS version, allowing users to run Large Language Models and Stable Diffusion models directly on their Android devices.

The app ensures complete privacy with no internet connection required and no data leaving the device.

Features

  • Offline AI: Run LLMs and Stable Diffusion models completely offline

  • Multiple Personalities: Choose from various character personalities

  • Custom Models: Load custom GGUF models tailored to your needs

  • Image Generation: Generate images using Stable Diffusion locally

  • Voice Support: Multiple text-to-speech voices available

  • Memory System: Long-term memory for personalized conversations

  • Roleplay: Engage in roleplay scenarios with AI characters

  • Task Management: Built-in task scheduling and reminder features

Solution

High level architecture of the application.

The Android version of Layla is built using Kotlin and follows modern Android development best practices. The app uses Jetpack Compose for a modern, declarative UI and implements the MVVM architecture pattern for clean code separation.

Room Database handles local data persistence, while Retrofit and OkHttp manage network operations when needed. Kotlin Coroutines ensure smooth asynchronous processing of AI models, and the app is optimized for devices with 8GB+ RAM to handle large language models efficiently.

Visit website

Technologies

Kotlin

Java

Android Studio

Android SDK

Gradle

Jetpack Compose

Material Design

Room Database

Retrofit

OkHttp

Coroutines

MVVM

More Applications

main*
Go Live