Concept2Comic

Website UI
Website UI in action
The Water Cycle with Party Cat
Water Cycle with Aqua
Party Cat's Pollen Adventure
Buzzy's Pollen Adventure

Inspiration

We wanted kids to have an early and easy introduction to AI in a family friendly way. These comics allow the students to learn about advanced topics through friendly cartoons where either fun gimmicky characters or our mascot, Party Cat, can introduce the topic in a comic strip.

What it does

Takes concepts from the user and explains them by creating a fun, easy-to-understand, comic panel format image using generative AI, where you can choose to have our mascot, Party Cat, explain it to you or another fun gimmicky character.

How we built it

We started building our project by setting up a team repositorie in GitHub and then messed around with prompt engineering with various AI. We used Gemini Flash 2.5 to test out different prompts, and while we ran into issues with setting up the Gemini API, and had to try solutions like running a local AI model, we were able to solve our issues and used Gemini API for text and image generation. To design our User Interface, we used a mix of JavaScript and CSS to get our desired look that promotes a cartoony family friendly look. We used Flask to power our back-end API.

Challenges we ran into

There were definitely a lot of challenges along the way. At first, we couldn’t get the Gemini API to set up properly, and when we tried to move our code to Replit, it didn’t work the way we wanted. The prompts we wrote for the AI also needed constant tweaking because they didn’t always give us good images. We experimented with different coding languages, but that just made things more confusing sometimes until we settled on JS and CSS. Eclipse and VSCode also gave us problems when we tried to set everything up. GitHub was another struggle because repositories can be tricky to manage as a team, especially because we were new to it. We had image generation bugs, UI customization didn’t always look right, and testing our code was frustrating and confusing for a long while. Even though we had to fight through all those problems, we managed to finish the project, and we’re proud of how it turned out! I couldn’t have asked for a better team and I’m excited with what we created!

Accomplishments that we're proud of

Accomplishments we’re proud of include getting the Gemini API to work, finalizing a prompt after serious turmoil, the fantastic look of our website, which was inspired by the 90ies comic look, and our amazing mascot, Party Cat!

What we learned

During this project, we picked up so many new skills. We learned how to use GitHub, which was confusing at first since most us were new to GitHub but eventually the repositorie made sense. We also figured out how to work with Replit and VSCode, and we got better at writing Python syntax. One of the coolest things we learned was prompt engineering, which is basically how to ask the AI questions in the smartest way. We also got to experiment with the Gemini API and Google AI Studio, which taught us a lot about how professional tools work. On top of that, we practiced teamwork, we worked through our doubts on different ideas for our project. We even learned how to use Node.js and Next.js, and we tested out the OpenAI API. Another big accomplishment was running a local AI model using Google Colab, which made us feel like real computer scientists, but we ended up not needing that anyway.

What's next for Concept2Comic

What’s next is creating a sign-up/log-in menu for user accounts, so that they can save their past comics in their own personal gallery and share them with friends and family. We also want to allow the user to generate comics in a custom style of their choosing, while also allowing users to create longer comics with more complex topics. We finally hope to speed up image generation speed, so the user can get to their fun comics even faster!