Inspiration
College students in Bharat face intense academic and emotional pressure that often goes unnoticed. Exam stress, placement anxiety, CGPA comparisons, family expectations, and social isolation are common, especially in tier-2 and tier-3 colleges where access to mental health support is limited.
Many students hesitate to seek help due to stigma, cost, or lack of availability. Most support systems are also unavailable during the late hours when anxiety tends to peak. Clarity AI was inspired by the need for a private, always-available, and non-judgmental support system tailored to student life in Bharat.
What it does
Clarity AI is an AI-powered mental wellness companion designed to help college students manage emotional stress and academic anxiety.
It allows students to: Freely express thoughts and emotions through natural conversation. Receive empathetic, supportive responses focused on clarity and coping. Get personalized interactions using contextual memory. Discover nearby mental health professionals when needed. Access emergency support through a consent-based helpline connection in crisis situations. The system is designed to act as a first layer of emotional support, not a replacement for professional care.
How we built it
Clarity AI is built using a modular and privacy-conscious architecture. A mental health focused language model (MedGemma) runs locally using Ollama to ensure affordability and low-bandwidth usability. A Streamlit-based chat interface provides a simple and familiar user experience. Contextual memory stores basic user details with consent to personalize responses. User messages are monitored for crisis related keywords. On detecting high-risk situations, the system requests consent and then connects the user to an Indian suicide prevention helpline using Twilio. A search API is used to suggest nearby therapists or mental health professionals when requested.
Challenges we ran into
Designing empathetic responses without crossing into medical diagnosis. Implementing crisis escalation in a way that is ethical and consent-driven. Balancing personalization with user privacy. Building an India-first solution without relying on expensive cloud APIs.
Accomplishments that we're proud of
Building a fully functional AI mental wellness system rather than a demo or mockup. Implementing real-world emergency escalation with user consent. Designing a system that runs locally and is accessible in low-resource settings. Creating a student-focused solution aligned with real challenges in Bharat.
What we learned
Mental health applications require ethical design, not just technical accuracy. Students often need validation and clarity before advice. Consent and transparency are essential when AI interacts with vulnerable users. India-first AI solutions must account for cost, bandwidth, and accessibility.
What's next for Clarity AI
Planned improvements include: Multilingual and Hinglish support for wider accessibility. Academic stress specific modes for exams and placement preparation. Campus specific mental health resources and counselor listings. Offline-first deployments for rural and semi-urban colleges.
Log in or sign up for Devpost to join the conversation.