MY TALKIE - AI CHATBOT WITH VOICE ASSISTANCE

Abstract

Educational institutions repeatedly answer questions about admissions, examinations, attendance, fees, scholarships, placements, library services, and campus facilities. This paper presents My Talkie, a webbased AI chatbot with voice assistance and documentanalysis support for student queries. The application uses Python and Flask for request processing, HTML, CSS, and JavaScript for the interface, and the Groq API with the Llama-3.1-8B-Instant model for context-aware response generation. Students can enter questions as text or use browser speech recognition. Responses are displayed in the chat window and can be read aloud through speech synthesis. The system also accepts image, PDF, and text-file uploads: Tesseract OCR extracts text from images, pdfplumber reads PDFs, and text files are processed directly before the recovered content is supplied to the language model. The project architecture separates the user interface, Flask backend, fileprocessing services, AI service, and voice services. The documented executions demonstrate text-based question answering, voice interaction, image OCR, PDF explanation, text-file analysis, and college-oriented responses through a compact browser interface. The resulting prototype provides a practical foundation for continuous student assistance and can be extended with authentication, college ERP connectivity, multilingual interaction, chat history, cloud deployment, and personalized academic services.