Abstract: In today’s fast-paced digital world, personal assistants like Alexa or Google Assistant are becoming increasingly popular. However, most of these tools require constant internet access and send user data to cloud servers, which raises privacy concerns. Our project, AI Personal Assistant, aims to solve this issue by creating a fully offline, intelligent assistant that works locally on a user’s computer without depending on the internet. This assistant not only protects user data but also brings together artificial intelligence, automation, and IoT integration into one system. Developed using Python, the assistant features a modern and responsive graphical interface built with CustomTkinter and supports both voice and text commands. It uses the speech recognition for understanding user input and pyttsx3 for providing verbal responses. The integration of the Ollama 3.2 local AI model allows it to generate smart and relevant replies without any internet connection. For security, the assistant includes face recognition using OpenCV, ensuring that only the authorized user can access the system. In addition to software intelligence, the assistant also connects with real-world hardware using ESP32 and ESP8266 microcontrollers. These allow it to read data from sensors like temperature, humidity, motion, and GPS, effectively making it an IoT-enabled assistant. It can also perform tasks like sending WhatsApp messages, managing Google Calendar events, and automating YouTube or system applications using tools like pyautogui and pywhatkit. This project demonstrates how AI can be made more personal, private, and practical—offering a secure and smart solution. Keywords — AI Personal Assistant, Offline AI, Local AI Model, Voice Recognition, Text-to-Speech, Face Recognition, Python Automation, CustomTkinter, OpenCV, Speech Recognition