Cosmo Corral | Devpost

Inspiration

Our inspiration stems from the growing need to protect individuals, especially the elderly and vulnerable, from sophisticated online scams and the overwhelming digital distractions of the modern internet. We envisioned a "digital guardian" that could proactively identify threats like phishing and fraud in real-time while also helping users reclaim their focus from constant interruptions. The goal was to create a safety net that empowers users to navigate the digital world with confidence and peace of mind.

What it does

EyeCore is a comprehensive digital wellness and security platform. It operates through three core components:

A high-performance desktop client (written in Rust) that runs in the background, securely monitoring on-screen activity and system events to detect patterns associated with scams or distractions.
A powerful backend server (built with Python) that receives data from the client, applies machine learning models for advanced threat analysis, and manages user data.
An intuitive web dashboard (developed with Next.js and TypeScript) that provides users with real-time alerts, detailed insights into their digital habits, and a summary of detected risks, empowering them to take control of their online safety.

How we built it

We adopted a polyglot architecture to leverage the best technology for each part of the system:

Desktop Client: We chose Rust for its performance, safety, and low-level system access capabilities, making it ideal for efficient and secure data collection without impacting user experience.
Backend Server: Python was our choice for the server due to its robust machine learning ecosystem. We used it to build the data analysis pipeline, including the ML-powered scam detection classifier.
Web Dashboard: We used Next.js, TypeScript, and Tailwind CSS to create a modern, responsive, and user-friendly interface for data visualization and alert management.
Communication: Real-time communication between the client and server is handled efficiently via WebSockets.

Challenges we ran into

Integrating three distinct technology stacks (Rust, Python, and TypeScript/Next.js) presented a significant challenge, requiring careful API design and data schema management to ensure seamless communication. Developing the Rust client to perform cross-platform system monitoring was complex, as was training an accurate machine learning model for scam detection with limited data. Ensuring the entire system was performant and secure, especially when handling sensitive user data, was a constant focus and a major hurdle we worked hard to overcome.

Accomplishments that we're proud of

We are incredibly proud of building a functional end-to-end system that integrates multiple complex technologies. Creating a performant Rust client for real-time data collection and a Python backend with a working machine learning model for threat detection are major technical achievements. The result is a cohesive platform that successfully identifies potential risks and provides actionable insights through a polished and intuitive dashboard.

What we learned

This project was a deep dive into full-stack, polyglot development. We learned how to effectively integrate Rust, Python, and TypeScript into a single, cohesive application. We gained valuable experience in system-level programming with Rust, building and deploying machine learning models, and designing scalable, real-time web applications. Most importantly, we learned how to tackle a complex social problem with a multi-faceted technological solution.

What's next for Cosmo-Corral

Our vision for EyeCore is to expand its capabilities as a comprehensive digital guardian. The next steps include:

Enhancing the ML Model: We plan to refine our machine learning classifier with more diverse datasets to improve its accuracy and reduce false positives.
Expanding Platform Support: We aim to develop client versions for macOS and Linux to reach a broader user base.
Adding More Features: Future features include advanced parental controls, productivity tools to manage digital distractions, and deeper integration with web browsers for more granular content analysis.
Mobile Application: A companion mobile app is on the roadmap to provide users with on-the-go alerts and access to their dashboard.