AdLibGPT

Inspiration

Our team was motivated to develop a sophisticated app that could harness the full potential of OpenAI's Language Models through their OpenAI API. We recognize that Large Language Models are the future and thus, we set out to explore its possibilities.

What it does

Our application provides the user with a selection of story themes, generating a custom ad-lib story based on their preference. The user fills in the blanks of the ad-lib, after which they can submit it to find out the ending of their story. This is for the SpaceOps track.

How we built it

Our team utilized a range of full stack web development tools, including React and Javascript for the front-end, and Express, Javascript, and the OpenAI API for the back-end REST API integration.

Challenges we ran into

We encountered difficulties in ensuring the consistency of output format due to the Large Language Models' (LLMs) non-deterministic nature. To overcome this issue, we employed prompt engineering methods to guide the model toward generating consistent format outputs.

Accomplishments that we're proud of

We take pride in leveraging OpenAI's latest technology and introducing a fun twist to our project. Our team successfully demonstrated the ability to create something exceptional with minimal code by leveraging the API.

What we learned

Throughout the project development, we gained valuable insights into prompt engineering practices and discovered the immense power of the tool with relatively little code.

What's next for AdLibGPT

Our team looks forward to expanding the capabilities of AdLibGPT and deploying it as a fun game for all to enjoy. We plan to include more specific and customizable story themes to offer a more personalized user experience.

Built With

• express.js
• javascript
• node.js
• openai
• react
• vite