Storyteller AI

Inspiration

We believe that a brighter future lies in the hands of the bright minds of the children of today. As adults, we might be changing today but kids are the tomorrow of humanity. For a bright future ahead of us we need bright minds to walk the path that lies ahead of the world. Sadly, the content that the newer generation has been exposed to recently has been extremely toxic and repetitive making a toxically addictive environment for the young ones while hindering their mental development. As many of our own related cousins and siblings already suffer from the same fate as most children exposed to the internet today, we felt compelled to bring a change for the betterment of our future. We place our values in the importance of a curious, creative and educated mind and to ensure that our future generation has an environment to promote such development, we came up with Storyteller AI.

What it does

Our AI aims to provide a very comfortable space for children of any age to learn something through an interactive story ranging over 20+ genres and over 15 types of emotions. The AI is very easy to use with little to no user input required and produces very creative and engaging outputs to make sure that the users don't lose any sense of interest. The AI takes text input from the user and responds in speech as well as text with a humane AI voice model. It frequently asks you regarding the course of the story and your opinion on the plot so far and it uses that data to generate further parts of the story with your feedbacks and preferences in mind. It ensures friendly, comfortable, inspiring, and safe conditions for the user while being very creative and respectful at the same time. Our Storyteller ensures a safe and easy to understand speech at all times while maintaining an engaging and instant-generative story. We are proud of its reliability and helpfulness.

How we built it

We first laid out the plan for the project and decided what APIs to use for different parts of our project. After researching more about the sponsors, we decided the most suitable APIs for our workload. We decided to use OpenAI API as the basic chatbot training model, Gradio for the front-end, Elevenlabs for text-to-speech and the RAFT method for prompt engineering. After distributing the workload amongst the team members, we made sure to communicate with each other to ensure the reliability of each and every step.

Challenges we ran into

We are CS majors so we don't know how to communicate with each other This was our first hackathon so we barely had any idea of what the 'hackathon' means Some sponsor APIs were very very very hard to integrate into our project so we spent hours and all of our mental health trying to make something work which we didn't need. Sleepy Syntax because computer science

Accomplishments that we're proud of

As a team, we are proud of making it through this hackathon, as our team is participating in one for the first time, considering it is such a big platform to meet new people and learn everything from the smartest people of our industry. We are also proud of creating our first AI project together considering this is our first time working together. Most of all we are really proud of how suitable our Storyteller is for children as no matter the amount of improper language or scenarios we try to push onto the AI (for research purposes) it ensures to teach the flaw in that train of thought with a friendly, interactive and very easy to understand examples. And most of all, we are proud of all of our teammates for making this possible.

What we learned

We learned a lot more than we were ready for and all of it was very exciting information that will change our future and career for the better. We learned about CoT for AI models, application of RAFT, how to integrate AWS, Groq, Hume ai, You.com and many more amazing AI company models, we learned a lot from our competitors as well as their experience was very evident in their polished and palatable projects. Most importantly we suppose, we learned to fall in love with the process as the journey follows along with it.

What's next for Storyteller AI

Storyteller AI has limitless opportunities and we plan on making the best out of them. We are further going to implement Speech-to-Speech for even easier access to children. We plan on a big update later on in the future where we integrate Hume AI into our project and analyze the emotions of the user on the basis of speech interactions and generate the story to elevate any negative feelings so that the user feels more comfortable, understood and connected. Finally we plan on integrating image generation for potential reference to make the story more engaging for the user.