Privacy Agent

Overview

This is a full-stack AI-Driven Mobile Assistant with On-Device Memory. I noticed that many chatbots save user data on the cloud. If an attack happens on the cloud, sensitive personal information can be exposed. To solve this, I developed this chatbot which stores the user's chat and AI memory locally on the device for complete privacy.

See Project:https://privacy-agent-app-download.vercel.app

Workflow

The application uses React Native for the user interface and a Node.js/Express backend to call the Google Gemini API. I utilized SQLite to store data locally on the device. I also added a useful Automatic Notification Schedule feature: you can ask AI to remind you to do a task at a specific time, and the device will notify you locally. Furthermore, the app supports Session-Based Chats so users can organize different chats.

Problem & Solution

Analyzing the current state of AI assistants:

Problems:

Most AI assistants rely on cloud processing, requiring the upload of sensitive personal data and chat history.
This creates a high risk of data leakage and privacy breaches since users have little control once data leaves the device.

Solution:

Privacy Agent provides a 100% local-first memory system using SQLite, ensuring chat history and reminders remain on the device.
By integrating Google Gemini 2.5 Flash-Lite and Gemma 3 27B via a robust Node.js backend, the app delivers high-speed, complex AI reasoning without compromising user privacy.

I am interested in backend development and continuously learning about it. If you find a backend developer then feel free to ask.

Quick Links

Home Skills Experience Projects

Contact On

+91 9429248465

jayjoshi1912007@gmail.com

@jayjoshi19