This project has been conceived as an extension of a homework assignment from the AI_devs course. The system is presented in a user-friendly console application format. The primary goal of this venture is to provide a playground for my experimentation revolving around technology fuelled by LLM models and the OpenAI API.
- Users have the ability to create articles employing the LLM model simply by supplying an article title.
- Article creation unfolds in the following manner:
- The LLM model generates a list of chapters based on the given title.
- These chapters are then stored in a database.
- The LLM model confers a unique content and summary for each chapter, one at a time.
- Succeeding chapters are formulated from the titles and short summaries of preceding chapters - enhancing contextual connectivity.
- Upon the completion of all chapters, the article is stamped as finished.
- Users can abort the article generation process at any given time.
- Users have the ability to choose an incomplete article and resume its generation from the last completed chapter.
- Users can access a comprehensive list of all articles, finished or otherwise.
- Users are given the option to delete articles.
- Familiarised self with Prisma controller testing.
- Acquired skills in inline prompting.
- Gained proficiency in incorporation of GPT models using the OpenAI API.
- Mastered Prisma mocking techniques.
- Prisma
- OpenAI API
- GPT Models
- Axios
- Node.js
- Jest
- TypeScript