Εφαρμογή Android για ιδιωτική αλληλεπίδραση με LLMs σε τοπικό περιβάλλον
Android application for private interaction LLMs in a local environment
View/
Keywords
Εφαρμογή android ; Προσωπικός βοηθός ; Ιδιωτικότητα ; Τοπική λύση ; Μεγάλα γλωσσικά μοντέλα ; Android application ; Personal assistant ; Privacy ; Local solution ; Large language modelsAbstract
The present thesis focuses on the development of an application that allows users to interact with Large Language Models (LLMs) without the use of cloud services, offering greater security and control over their data. The application, called "Home AI," utilizes resources from the user's local computer, providing personalized responses.
The main objectives of the application include data privacy through local processing, personalized responses based on the user's data, creating a user-friendly environment, and leveraging local resources to eliminate the need for external services.
To achieve these objectives, a client-server architecture system was designed, providing a secure user experience. The server is based on Role-Based Access Control (RBAC) and implemented using modern technologies such as FastAPI, security with JWT, Postgres, ChromaDB, and Ollama for interaction with LLMs.
The client is an Android application that allows interaction with the server through a friendly and functional User Interface (UI). The client development was based on Clean Architecture, utilizing technologies such as Retrofit, OkHttp, Jetpack Compose, and Kotlin Coroutines, ensuring a flexible user experience.
Department
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
44Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα