SkyScraper

🌟 Inspiration

The aviation industry operates under immense pressure to balance safety, efficiency, and customer service in real time. From unpredictable weather patterns to high ATC load, risk management can feel reactive rather than proactive. We were inspired to build SkyScraper as a modern AI-driven solution that gives operators, airlines, and even customers real-time insights, proactive alerts, and automated assistance — helping them make safer, faster, and smarter decisions.

🚀 What it does

SkyScraper is a real-time aviation safety and operations dashboard combined with AI assistants. ✅ It provides live flight tracking with detailed risk scores based on weather, ATC load, aircraft type, and more. ✅ It features an AI chat agent that answers safety-related queries, recommends rescheduling, and evaluates flight risks. ✅ It integrates a Vapi-powered voice agent for lead qualification, smart calling, and customer follow-ups, enabling human-like conversations over the phone. ✅ The platform issues automated alerts when flights surpass critical risk thresholds, helping operators act fast.

🛠 How we built it

Frontend: Next.js + TailwindCSS for a clean, responsive UI with interactive components (scroll areas, dialogs, badges). Flight data: Mock + live API integrations (e.g., AviationStack) for simulating flight risk metrics. AI agents: Custom AI chat assistant using Next.js API routes + OpenAI API for responses. Vapi integration to handle outbound calls and lead qualification, with webhooks to capture results. State management: React hooks and context for managing flight data, messages, and agent states. Deployment-ready: Designed with scalability in mind, easily hosted on platforms like Vercel. ⚡ Challenges we ran into

Real-time simulation: Since we didn’t always have access to live flight data during development, we had to design realistic mock data structures that could seamlessly integrate with actual APIs later. Vapi integration: Handling phone interactions and lead qualification logic via Vapi while ensuring responses sync back into the dashboard required careful API design and webhook handling. Designing meaningful risk scores: Creating a risk model that feels realistic and actionable, even in demo scenarios, took iteration. Agent communication UX: Balancing between text and voice agents so they enhance — not overwhelm — the user experience. 🏆 Accomplishments that we're proud of

Building an end-to-end aviation safety system that combines data, AI, and automation. Seamlessly integrating chat and voice agents into a single dashboard. Creating a clean, user-friendly interface that surfaces critical flight information clearly. Designing a flexible architecture that can easily hook into real APIs or additional data sources. 📚 What we learned

How to design a system that mixes real-time data, AI text, and voice automation cohesively. The importance of clear communication UX when users interact with multiple AI agents. How to work with external APIs like Vapi for phone automation and manage async events via webhooks. The value of designing scalable data structures early, so adding live data sources or features later becomes trivial. 🔮 What's next for SkyScraper

Hook up real-time flight data sources (e.g., FAA feeds, OpenSky API) for production-grade accuracy. Add user authentication + role-based access (e.g., ATC, airline ops, customer view). Extend the Vapi agent’s capabilities to handle more complex customer interactions, including multi-turn conversations for bookings and cancellations. Build analytics dashboards for historical flight risk patterns and operational insights. Explore mobile-first design for operations teams on the go.

Built With

• claudesonnet4
• google-cloud
• groq
• nextjs
• vapi