Design of text generator application with OpenAI GPT-3

Abstract

The increasing need for text content creation today challenges the development of systems that can alleviate the need for text creation. Currently, text generation is done manually and has various shortcomings, especially in terms of time constraints, human error, limited creativity, and writing that tends to be repetitive by certain people, which can cause a decrease in quality and diversity in the sentences produced. This research was conducted by designing an AI-based text generator application using the GPT-3 language model to generate text automatically and help overcome some obstacles. Applying this app will increase efficiency and productivity, increase the writer's ideas and creativity, automate routine tasks, and produce exciting and communicative sentences. The app's ability to generate text quickly and accurately and be personalized makes it valuable in various fields. The method used in this research is implementing the GPT-3 language model APIs into the text generator application created so that the application can connect with the GPT-3 engine that has been modified in its prompting method. The output of this application is a text that has been adjusted to the user's needs through keywords entered on the web interface system. The result is that the text generator application is good enough to be implemented in various fields, especially text content generation.