Inspiration
We wanted to create a tool that could genuinely help people in moments of mental clutter - not just track tasks, but truly understand how you're feeling, why you're stuck, and guide you back to clarity. NeuroTwin was born from the idea of giving every person a real-time, emotionally aware cognitive mirror.
How we built it
We used:
Vapi for voice journaling and input DeepMind for emotional tone and intent analysis LlamaIndex to retrieve context from past reflections Orkes to trigger real-world interventions like timers or task updates Mixpanel to track feedback and outcomes HockeyStack, Datadog, and ZeroEntropy for behavior tracking and optimization
Challenges we ran into
Coordinating real-time agent flows within a short time frame. Balancing emotional intelligence with actionable output — not sounding robotic.
Accomplishments that we're proud of
Completed a full voice-to-action pipeline with emotion detection and tailored nudges. Seamlessly integrated multiple sponsor tools — each with a meaningful role. Built a polished, demo-ready prototype that actually feels helpful and human.
What we learned
Voice is one of the most intuitive ways to surface mental friction - and when paired with emotional intelligence, it unlocks powerful real-time support. MCP-style agent collaboration can be lightweight yet incredibly effective. Good UX in mental health tools means timing, tone, and trust - not just accuracy.
What's next for NeuroTwin
Integrate deeper journaling memory and feedback loops using vector stores. Enable proactive check-ins and AI-led micro-therapy sessions. Expand into personalized coaching for focus, burnout, and emotional regulation.
Log in or sign up for Devpost to join the conversation.