Abstract
This project presents the development of a mental health chatbot web, and mobile application to provide emotional support, tailored advice, and guidance on potential user mental health issues. Religious quotes from a variety of sacred texts across multiple religions are also implemented to provide comfort to religious users. This model incorporates several natural language processing techniques (NLP), and large language models (LLM’s) to generate empathetic, relevant, and supportive responses to user queries. This model is grounded on a curated dataset of mental health conversations to retrieve relevant data using TF-IDF vectorization and cosine similarity. The web application is built using Streamlit and hosts the model on a browser. The mobile application is built using Swift and incorporates a Rest API to connect the mobile application to the backend logic via Uvicorn and HTTP requests. The model is evaluated on performance metrics such as latency, and emotional relevance of the generated responses.
Advisor
Musgrave, John
Department
Computer Science
Recommended Citation
Headen, David Jerome Jr., "Echo: Design and Implementation of an AI-Powered Web & Mobile Mental Health Application Using Retrieval-Augmented Generation" (2026). Senior Independent Study Theses. Paper 13291.
https://openworks.wooster.edu/independentstudy/13291
Disciplines
Physical Sciences and Mathematics
Publication Date
2026
Degree Granted
Bachelor of Arts
Document Type
Senior Independent Study Thesis
© Copyright 2026 David Jerome Headen Jr.