Reflect is a cutting-edge mental health counseling platform designed to bridge the gap between students and counselors. By integrating real-time video therapy with advanced AI analysis, mood tracking, and personalized care plans, Reflect provides a holistic approach to mental well-being.
We are proud to share that this project earned the top spot at the KSU AI Hackathon 2025.
This project was built by:
- Aaryan Kapoor - Lead + Solutions Architect + AI Engineer
- Eyasu Smieja - Researcher + FastAPI Developer
- Chris Martinez - Researcher + Frontend Designer
- Evan Valencia - Frontend Designer
Our video calling system is powered by Gemini 3, enabling native video understanding that extends far beyond basic transcription. It interprets visual cues, tone shifts, and contextual signals - effectively seeing what happens between the lines. After each session, you receive a comprehensive AI-generated summary that highlights key moments, meaningful insights, and notable video-based observations.
Comprehensive views for both counselors and students to track progress, mood trends, and upcoming sessions.
| Counselor Dashboard | Student Dashboard |
|---|---|
![]() |
![]() |
| Monitor student trends, AI summaries, and risk alerts. | Track mood history, access journals, and view upcoming sessions. |
Leverage Google Gemini to generate personalized pre- and post-session questionnaires based on student history and session context.
| Step 1: AI Prompt | Step 2: Generated Questions | Step 3: Student Response |
|---|---|---|
![]() |
![]() |
![]() |
| Counselors prompt AI with context. | AI generates tailored questions. | Students engage with personalized forms. |
Empower students to document their thoughts and follow structured growth plans.
| Journaling | Growth Plan |
|---|---|
![]() |
![]() |
| Private or shared entries for self-reflection. | Structured tasks and goals for continuous improvement. (Assigned by Counselor) |
Counselors can assign therapeutic tasks and track completion.
Assign and monitor therapeutic homework.
Reflect is built on a robust, modern stack:
- Frontend: React (Vite/CRA), Clerk (Auth), Convex (Real-time DB)
- Signaling Server: Node.js, Socket.IO (WebRTC coordination)
- AI Service: Python, FastAPI, Google Gemini (Analysis & Content Generation)
For a deep dive into the system architecture, check out the Architecture Documentation.
- Node.js 18+
- Python 3.10+
- Google Gemini API Key
-
Clone the repository
git clone https://github.com/Aaryan-Kapoor/Reflect-KSU-Hackathon.git cd Reflect-KSU-Hackathon -
Install Dependencies
# Root (Signaling Server) npm install # Client (Frontend) cd client npm install cd .. # AI Server python -m venv venv source venv/bin/activate # or venv\Scripts\activate on Windows pip install -r requirements.txt
-
Environment Setup Set your Gemini API key:
export GEMINI_API_KEY="your-api-key-here"
You can run all services using the helper script:
./run_all.shOr run them individually in separate terminals:
-
Signaling Server (Port 5000)
npm start
-
AI Server (Port 8000)
source venv/bin/activate python ai_server.py -
Frontend (Port 3000)
cd client npm start
This project is licensed under the MIT License.
Built for the KSU AI Hackathon 2025.






