INSPIRATION
Today’s youth face rising levels of depression, anxiety, and loneliness. Many people don’t feel safe talking openly about their struggles, which is why suicide rates are alarmingly high. We wanted to create a safe, anonymous space where anyone can talk without fear of judgment — while still ensuring that silent cries for help don’t go unnoticed. That vision inspired SilentSupport.
WHAT IT DOES
SilentSupport is a GenAI-powered emotional companion that chats like a friend. It detects emotions, provides quick relief tools (breathing, grounding, affirmations), and offers both voice and text chat. If the system senses severe distress or suicidal intent, it silently alerts nearby certified professionals — without breaking user trust. Users can also manually request help at any time.
HOW WE BUILT IT
Frontend (Next.js + React + TailwindCSS) for a soothing blue-white interface, with animations, mic-based voice chat, and quick-help grids. Backend (Node.js + Express) with APIs for chat, voice, alerts, and professional dashboards. Supabase (Postgres + PostGIS) to store users, chats, sessions, and location-based professional atching. Cohere API for chat generation and emotion-aware responses. Web Speech API (speech-to-text + text-to-speech) for real-time voice interaction. EmailJS + Nodemailer for sending crisis alerts to professionals and guardians.
CHALLENGES WE RAN INTO
Ensuring privacy while enabling silent crisis detection was tricky — we balanced thresholds to avoid false positives. Integrating real-time voice chat with text and emotion detection smoothly. Designing a friendly but professional UI that feels safe and trustworthy. Database design for both anonymous sessions and registered users with guardians and professionals. Accomplishments that we’re proud of Created a working AI companion that can genuinely uplift conversations. Built crisis detection logic that saves lives without alarming users unnecessarily. Designed a professional dashboard for mental health workers to view cases and respond Developed quick-relief tools (breathing, mindfulness, grounding) that make the app more than just a chatbot.
WHAT WE LEARNED
How to combine GenAI with real-world safety mechanisms. The importance of ethical design in AI mental health applications. Building a production-ready stack with Supabase, Cohere API, and Next.js. What’s next for SilentSupport Adding multilingual support for global accessibility. Expanding the network of verified professionals. Mobile app version for 24/7 availability. Research partnerships with mental health organizations.
Built With
- cohere
- emailjs
- express.js
- next.js
- node.js
- nodemailer
- react
- responses
- supabase
- tailwindcss
- webspeech
Log in or sign up for Devpost to join the conversation.