Aura AI: Your Proactive Partner Against Burnout

The Inspiration

Burnout doesn't shout; it whispers. By the time most people realize they are drowning, they lack the mental energy to reorganize their lives. Aura AI was built to catch you before the fall, a voice-first assistant that doesn't just listen to your words, but hears your stress.

What It Does

Aura uses Real-Time Voice Prosody to detect sub-textual signals like tension, fatigue, and anxiety.

• Empathic Listening: Analyzes 48 emotional signals in your speech.
• Autonomous Action: If you say "I can't handle this," Aura doesn't just sympathize, it triggers a manage_burnout tool to reschedule low-priority tasks and clear your calendar.
• Visual Feedback: A glassmorphic "Zen Orb" shifts from Calm Teal to Stressed Rose, visualizing your mental load in real-time. The Tech Stack
• Frontend: Next.js 15, TypeScript, Framer Motion (for "morphing" task transitions).
• AI Core: Hume AI Empathic Voice Interface (EVI) for prosody analysis.
• State: Zustand with persistence to track stress trends over time.
• Logic: A custom Weighted Stress Algorithm that prioritizes signals like distress and fatigue over neutral tones.

Key Challenges & Solutions

• The "Optimism" Bias: Early versions let mild calm signals override high stress. Solution: Implemented Conditional Damping, reducing the weight of "calm" signals by 80% when stress markers are detected.
• LLM Hallucinations: AI sometimes uses "done" instead of "complete." Solution: Built a Normalization Layer to map natural language to strict system actions.
• Sync Lag: Updating task lists via WebSockets can be heavy. Solution: State-tracking to ensure updates only fire when tasks actually change.

Accomplishments

• True Autonomy: Aura makes executive decisions (task rescheduling) based on emotional state, not just manual commands.
• Subtle Accuracy: Successfully identifies elevated stress even when the speaker's words remain professional or neutral.
• Zen Aesthetic: A cohesive UI that reduces cognitive load through soft gradients and glassmorphism.

What’s Next?

• Predictive Warning: Identifying burnout patterns before the week begins.
• Deep Integration: Connecting to Google Calendar/Outlook to move actual meetings.
• Team Delegation: Detecting when a user is overwhelmed and suggesting tasks to delegate to available colleagues.

Built With

• antigravity
• gemini
• hume-ai
• next-js