--- title: Mental Health Support Assistant emoji: 💚 colorFrom: green colorTo: blue sdk: gradio sdk_version: 4.44.0 app_file: app.py pinned: false license: mit --- # Mental Health Support Assistant 💚 A compassionate AI assistant that provides mental health support using Retrieval-Augmented Generation (RAG). This application combines semantic search with AI generation to offer contextually relevant, empathetic responses based on mental health counseling conversation data. ## Features - **Empathetic Responses**: Provides supportive, understanding responses to mental health concerns - **RAG-Powered**: Uses vector search to find relevant context from counseling conversations - **Safety-First**: Includes crisis detection and appropriate resource recommendations - **Professional Boundaries**: Clearly states limitations and encourages professional help when needed ## How It Works 1. **Data Processing**: Mental health counseling conversations are cleaned and embedded using sentence transformers 2. **Vector Search**: User queries are matched against the embedded conversation database 3. **Response Generation**: OpenAI's GPT model generates empathetic responses based on retrieved context 4. **Safety Measures**: Built-in crisis detection and resource recommendations ## Important Disclaimer ⚠️ This is an AI assistant for informational and emotional support purposes only. It is not a substitute for professional mental health care, therapy, or medical advice. If you're experiencing a mental health crisis, please contact emergency services or a crisis hotline immediately. ## Crisis Resources - **National Suicide Prevention Lifeline:** 988 (US) - **Crisis Text Line:** Text HOME to 741741 - **International:** [Find local crisis centers](https://www.iasp.info/resources/Crisis_Centres/) ## Setup 1. Set your OpenAI API key in the environment variables or Hugging Face Spaces secrets 2. Optionally set your Hugging Face token for dataset access 3. The application will automatically download and process the mental health dataset on first run ## Technology Stack - **Frontend**: Gradio - **Vector Database**: ChromaDB - **Embeddings**: Sentence Transformers (all-MiniLM-L6-v2) - **LLM**: OpenAI GPT-3.5-turbo - **Dataset**: Mental Health Counseling Conversations