The number of content creators in the cyber world is growing faster year by year, and the competition fiercer. Large video platforms such as YouTube offers creators incentives to upload original content more frequently. However, every creator has a different definition of novelty and uniqueness. The biggest challenge a creator has to face every day lies in the generation and practice of ideas. As a result, a customized and efficient “idea” generator has become necessary in our times, and any content creator, whether video, advertising, or writing, can benefit from making their content unique efficiently without losing their style. The advent of GPT-2/3 makes this possible, and in this thesis, I will explore the types of models, the feasibility of streamlining, and the practical challenges of customizing a text generator for content creators nowadays.