ScriptGenie AI: A Groq API and LLaMA 3-Based Full-Stack Framework for Automated YouTube Script Generation
Kumar Harsh, MD Tayyeb Hussain, Satyam Kumar, Jayanta Kumar Bag, T. K. Mukherjee · Zenodo (CERN European Organization for Nuclear Research) · 2026
ScriptGenie AI: A Groq API and LLaMA 3-Based Full-Stack Framework for Automated YouTube Script Generation 1Kumar Harsh, 2MD Tayyeb Hussain, 3Satyam Kumar, 4Jayanta Kumar Bag, 5Tilak Mukherjee 1234Dept. of Electronics & Communication Engineering, Haldia Institute of Technology, Haldia, Purba Medinipur, West Bengal, India 1UG Student, Department of ECE, Haldia Institute of Technology, Haldia, West Bengal 2Assistant Professor, Department of ECE, Haldia Institute of Technology, West Bengal, India ABSTRACT Scripting remains one of the most time-consuming stages involving YouTube video production, requiring creators to structure a hook, introduction, main content, and call-to-action while also drafting titles, tags, and descriptions for discoverability. ScriptGenie AI is a full-stack web application that automates this process by converting a single user-supplied topic into a complete, structured video script. The system integrates a vanilla HTML/CSS/JavaScript front end with a Node.js and Express.js back end, which communicates with the Groq inference API so as to query the open-weight LLaMA 3 language model through a structured, JSON-constrained prompt. The present paper details the system architecture, the rationale for selecting an inference-API-based design over local model hosting, the request-response workflow, and the prompt-engineering strategy used to obtain reliably structured output.