DayLight

Inspiration

The three of us kept noticing the same pattern in our busy lives. We'd hit burnout, feel mentally drained, and each handle it differently. Some of us would take walks, others would grab coffee or take quick naps. But the common thread was clear: we all sought out specific activities that refreshed us and balanced our days. That's when the concept of cognitive load points clicked. What if we could quantify mental energy the same way we budget time? What if your calendar could tell you "hey, you're at 18 out of 20 points today, maybe skip that optional meeting"? DayLight was born from wanting to prevent burnout before it happens, not just react to it.

What it does

DayLight introduces Cognitive Load Budgeting to calendar management. It assigns cognitive costs to your calendar events based on duration, participant count, whether there's an agenda, context switching, and scheduling patterns. You start each day with a 20-point budget.

The app monitors your actual stress levels during work sessions using webcam-based vital signs analysis:

Pulse rate from facial blood flow
Breathing rate detection
HRV (Heart Rate Variability) as the primary stress indicator
Real-time stress classification

When you're overdrafted, it suggests concrete recovery activities and can even optimize your entire week by redistributing movable events to balance your cognitive load. It integrates with Google Calendar and uses AI to automatically classify events as meetings, deep work, admin tasks, or recovery time.

How we built it

We built DayLight with a React 18 frontend connected to a FastAPI Python backend via WebSockets for real-time vital signs streaming. The most complex piece was integrating the Presage SmartSpectra SDK, a C++ library for extracting vital signs from webcam video.

Our tech stack:

Frontend: React 18, Tailwind CSS, Recharts
Backend: FastAPI (Python), WebSocket streaming
Vital Signs: Presage SmartSpectra SDK (C++), Docker containerization
AI: Gemini API for event classification

We containerized the entire Presage daemon using Docker and established a TCP connection between it and our Python backend, which then streams processed metrics to the frontend over WebSockets. All of this came together in 24 hours at SpartaHack 2026.

Challenges we ran into

Getting the Presage SmartSpectra SDK working was brutal 😅. It's cutting-edge technology still in early development stages, and it only runs on Ubuntu. We spent hours figuring out:

How to properly host the service
Establish stable WebSocket connections
Bundle everything into a Docker container that could communicate with our backend

The multi-layer architecture with video streaming from frontend to backend, TCP connections to the Presage daemon, and metrics streaming back through WebSockets had so many potential failure points. Debugging connection issues across three different technologies while racing against the 24-hour clock was intense.

Accomplishments that we're proud of

WebSocket magic: Handled concurrent connections flawlessly. Multiple users could run Sage sessions simultaneously without issues ✨
Smart algorithms: Our event classification and cognitive load point assignment felt natural and intuitive, not forced or gamified
Learn-on-the-fly: One of our team members learned React completely from scratch during the hackathon and was able to add functional features on the go

What we learned

This hackathon was a masterclass in leveraging AI assistants for coding. The productivity boost was unreal and changed how we approach development. We learned how to integrate and containerize third-party services, specifically working with a C++ SDK in a Docker environment while maintaining WebSocket communication across the stack.

Key takeaways:

Building with cutting-edge, early-stage technology teaches resilience and creative problem-solving when documentation is sparse
AI coding assistants are game-changers for productivity
You can build something genuinely useful in 24 hours when you're solving a problem you personally experience every day

What's next for DayLight

Calendar Integration:

Expand beyond Google Calendar to include Outlook, Apple Calendar, and other platforms
Fix rough edges in our current Google Calendar implementation

Vital Signs Enhancement:

Work closely with Presage to better analyze metrics
Build a real-time metric evaluator service with lower latency

Wearable Integration:

Connect smartwatches and fitness trackers for holistic activity data throughout the day
Better understand what activities actually aid cognitive recovery for each individual

Intelligence: