DreamWeaver

Home page
Prompt page
Example story

Inspiration

We wanted to focus on education for children and knew the power of AI resources for both text and image generation. One of our members realized the lack of accessible picture books online and suggested that these stories could be generated instead. Our final idea was then DreamWeaver

What it does

Our project uses a collection of machine learning resources to generate an entire children's book after being given any prompt.

How we built it

Our website is built upon the React framework, using MaterialUI as a core design library.

Challenges we ran into

Being able to generate relevant images that fit the theme and art style of the book was probably our biggest challenge and area to improve on. Trying to ensure a seamless flow and continuity in the generated imagery was a valuable learning experience and something we struggled with. Another difficulty we faced was working with foreign APIs, as there were many times we struggled to understand how to access certain resources.

Accomplishments that we're proud of

Our use of machine learning, especially Cohere's LLM, is our project's most impressive aspect, as we combine different models and prompts to produce an entire book for the user.

What we learned

Our team worked extensively with AI APIs such as Cohere and Midjourney and learned a lot about how to use their APIs. Furthermore, we learned a lot about frontend development with React, as we spent a good amount of time designing the UI/UX of our web app.

What's next for DreamWeaver

Having more continuity between book images would be the first thing to work on for our project, as sometimes different characters can be introduced by accident. We are also looking into adding support for multiple languages.