Inspiration
Even highly technical people fall for phishing, spoofed domains, and social engineering. Experience does not eliminate risk. Attackers exploit urgency, trust, and human behavior more than technical weakness. We were inspired by a simple observation: security tools often protect data and are very fragmented, with a lack of communication with each other, causing multiple pop-ups, missing password information, and adding to the overall device vulnerability by increasing the attack surface. This app is designed to bring everything together: encrypted vaulting, phishing/domain checks, email verification, AI assistants, and accessibility read-aloud, so users get protection and guidance in one place.
What it does
GuardiaPass is a full AI security platform that combines encrypted credential management, browser protection, and intelligent assistants in one system. It secures passwords, identities, and sensitive media with strong local encryption, then extends protection into the browser with smart autofill, phishing/domain checks, breach intelligence, and suspicious-email filtering. It also includes multiple AI assistants: an in-app AI Security Advisor for ongoing guidance, extension-based page assistants for live site analysis, and contextual AI tools for password generation, strength analysis, and risk explanations. For accessibility, it can read selected webpage content aloud, helping users who prefer or require audio output. The project is designed to be practical and real-world: local-first secure storage, server-side API proxying to protect keys, real-time activity telemetry, domain/email trust checks, and assistant-driven workflows that help users make safer decisions faster. In short, GuardiaPass is an AI-powered security command center for everyday users: protect, verify, assist, and educate from one place.
How we built it
We built GuardiaPass as a React + TypeScript web app with a Chrome Manifest V3 extension and an Express backend. Security features were implemented with Web Crypto (AES-GCM + PBKDF2), local encrypted IndexedDB storage, and safer autofill/domain checks. AI and intelligence features were added through backend-proxied integrations (Gemini, Hunter, Backboard, ElevenLabs), then connected to UI modules for vault, advisor, breach checks, activity hub, email/domain filtering, and accessibility read-aloud.
Vibe Coding :-)
Challenges we ran into
Balancing strong security controls with real-world usability (especially autofill reliability on complex sites like Google using a web extension).
- Hardening cross-context extension messaging without breaking core flows.
- Making AI-based checks fast enough while still providing consistent, trustworthy output.
- Cross-platform startup/runtime differences (Windows vs macOS/Linux process behavior).
- set up application web extensions with high-level security while still allowing it to communicate encryptedly with the application.
What we learned
Strong encryption alone does not change behavior. Clear explanations without technical depth do not create real protection. Effective security requires both mathematical rigor and intuitive guidance. We also learned that key management, recovery logic, and user experience design are just as important as cryptographic strength.
What's next for GUARDIAPASS
Next, we plan to expand browser support, improve real-time threat intelligence, and refine the AI advisor with deeper contextual awareness. We aim to enhance enterprise-ready features, improve performance under heavy usage, and strengthen automated phishing detection models. Our long term direction is to evolve GuardiaPass from a secure vault with an advisor into a fully adaptive personal security layer that operates seamlessly across devices and environments.
Built With
- backboard.io
- chrome
- css
- elevenlabs
- gemini
- hibp
- html
- hunter.io
- indexeddb
- javascript
- lucide-react
- node.js
- npm
- postgresql
- react
- recharts
- sql
- tailwind
- typescript
- vite
- web
Log in or sign up for Devpost to join the conversation.