Flow State

main
flow state interface
flow state idea generation space
flow state dna refer to demo video

Inspiration

FlowState was inspired by the elusive concept of “flow”—that peak mental zone where focus sharpens, distractions fall away, and time seems to dissolve. We don’t believe creativity or deep concentration should be reserved for lucky moments or rare sparks of genius. Instead, we set out to biohack our ability to intentionally activate that state. FlowState aims to bridge biological feedback with productivity, creating a dashboard that visualizes study intensity by directly correlating it with real-time heart rate data.

How We Built It

Biometric Engine: At the core of FlowState is a custom browser-based biometric monitor. Using computer vision (photoplethysmography), we detect subtle facial flush changes through the webcam to calculate BPM in real time.

Backend: We used Python to process and smooth this raw data, then mapped BPM values to dynamic color gradients representing mental intensity.

Data Architecture: To manage session data, we built a JSON-based middleware layer backed by Firebase, allowing biometric sessions to be stored, retrieved, and analyzed cleanly.

Frontend: The interface was designed with Glassmorphism and soft, “gooey” animated backgrounds to evoke calm, fluid motion—reinforcing the idea that focus should feel organic, not rigid or stressful.

What We Learned

We learned how unpredictable and noisy biological data can be. A major focus became smoothing fluctuating sensor readings so the UI remained stable and visually coherent. We also gained deeper experience integrating AI (specifically the Gemini API) to transform nested JSON streams into CSV formats, making it easier to compare sessions and conduct post-analysis.

Challenges

Our biggest challenge was reducing latency and keeping the visuals tightly synchronized with heart rate changes. Achieving instant, lag-free color transitions required optimizing the entire pipeline from Python output to frontend rendering. We also iterated heavily on camera sensitivity to ensure accurate detection across different lighting environments and skin tones, refining the algorithm multiple times.