• Version
• Download 0
• File Size 324.00 KB
• File Count 1
• Create Date 17 February 2026
• Last Updated 17 February 2026

A GenAI-based Conversational CBT Therapist for Teenagers

Submitted by
Mr. Asvin Kant Sharma  : PRN No. 220105131116

Mrs. Pallavi Patil : PRN No. 220105131156
Mr. Vaibhav Suryawanshi : PRN No. 220105231063
Mr. Shreyas Jadhav :PRN No. 220105231077

Under the Guidance of
Prof. Dr. Sanjeev Shukla

Designation of Guide
Feb, 2026-27
Department of Computer Science and Engineering
School of Computer Sciences and Engineering
Sandip University Nashik

Abstract
1.Teenage mental health challenges such as stress, anxiety, loneliness, and emotional imbalance have been increasing at an alarming rate in recent years. Adolescents often struggle to articulate their emotions or seek timely help due to stigma, lack of accessibility, high consultation costs, and hesitation to speak openly about their struggles. As a result, many teenagers remain unsupported during critical phases of emotional development, which can lead to long-term psychological consequences. To address this gap, MindMeld is introduced as a Generative AI–based Conversational CBT (Cognitive Behavioral Therapy) assistant designed to provide real-time emotional support, early mental-health intervention, and guided CBT-based exercises in a safe and accessible manner.
2.The system integrates multiple advanced components, including Large Language Models (LLMs), Emotion Detection modules, CBT frameworks, and Context-aware Conversation Management. Together, these elements enable MindMeld to simulate empathetic, structured, and psychologically informed interactions. By analyzing user text or speech inputs, the system identifies emotional states and responds with CBT-inspired interventions such as journaling prompts, grounding techniques, thought reframing, and coping strategies. This ensures that teenagers receive personalized, context sensitive guidance that mirrors the therapeutic principles of CBT while maintaining a conversational and approachable style.
3.On the technical side, the backend leverages Natural Language Processing (NLP) pipelines, transformer-based models, sentiment analysis, and vector memory to preserve conversational context and continuity. Safety is prioritized through content-filtering mechanisms, crisis-detection modules, and escalation protocols that can guide users toward professional help when necessary. A user-friendly web interface further enhances accessibility, allowing teenagers to interact naturally with the system through both text and voice modalities.
4.This project demonstrates how Generative AI can be responsibly applied to create scalable, low-cost, and always available emotional wellness companions. While MindMeld is not intended to replace medical professionals or formal therapy, it offers a valuable supplementary tool for early intervention and emotional support. By combining cutting-edge AI technologies with evidence-based CBT practices, MindMeld represents a step toward democratizing mental health care and reducing barriers for teenagers worldwide.
5.Keywords: Mental Health, Generative AI, CBT, Emotion Detection, LLM, Teen Psychology