AI-Powered General Assistant Software

Abstract

This software names as AI-Powered General Assistant Software aims to be a stand- alone desktop assistant integrating Python, FastAPI, React, combining with both local (offline) and OpenAI-based models. The system provides a conversational interface that is capable of executing commands, managing files, browsing the web, reading on-screen content, and performing advanced reasoning tasks. Additionally, it has the ability to recognize face with DeepFace(python). The goal is to bridge the gap betweenAI’s accessibility and user autonomy. Unlike commercial assistants (e.g., Siri or Google Assistant), this system operates independently on the user’s machine, this ensures privacy and full control over users’ local conversation. The software has a very simple chat base frontend and a strong logical backend that support both offline AI (Ollama) and cloud AI (OpenAI API). It maintains sessions as memory with recall ability that hold the user and AI conversation in one page. By simply using command user can generation files (.txt, .docx, .pdf, .pptx) in their local storage. It can act on voice command by calling the wake word(customizable). This assistant software can also be used to organized the storage with in few seconds. It can also send email with any type attachment. By using the python Deepfake, it can organize images by face. This works best on organizing images from a google drive and send them to the person that matches by face and mail them to their personal mail.