Homepage
Characters
Chat UI
Chat Functionality

ImagiTalks combines nostalgia and artificial intelligence, allowing users to chat with the characters from their favourite cartoons. Cohere LLM skillfully mirrors characters' mannerisms, while ElevenLabs AI breathes life into their voices, elevating the immersive experience.

Inspiration

The memories of cartoons we adored in childhood.
The bittersweet feeling of finishing your beloved TV show.
The longing to get more heartfelt moments with the characters you love.

How we built it

Frontend Development Our journey began with the creation of a Figma prototype, streamlining the UI design process. Figma's user-friendly interface and versatile plugins facilitated a smooth design experience. After outlining the UI, including color palettes, logos, and fonts, we transitioned to building the frontend.

For optimal responsiveness, we chose a stack comprising React, Next, Tailwind, and Bootstrap. These technologies allowed us to create a visually appealing and adaptive web application.

Backend Development Simultaneously, on the backend, we established a PostgreSQL relational database to store our character models. We also have a Google cloud integration in the backend that allows us to handle audio file storage and conversion using cloud storage buckets.

We have implemented APIs for different operations in our application such as the ability of character persona recreation and speech.

The backend was built in Flask and leveraged a lot of open source libraries to manage dependencies.

AI Cohere AI played a pivotal role in character emulation on the backend. This technology allowed us to analyze user inputs, generate responses, and maintain coherent dialogue flow, enhancing the natural language processing capabilities of ImagiTalk.

The Elevenlabs library seamlessly converted text responses from Cohere AI into expressive character-specific voices, adding an immersive layer to the interaction.

Challenges we ran into

Integrating Auth0's authentication into our frontend
CORS
One of the challenges we encountered was setting up Google Dev Platform. The complexity and extensive setup process posed a significant hurdle for our team.

Accomplishments that we're proud of

Making a functioning product (that is also actually fun to use :)) despite all of the obstacles.
Developing a full-stack app that utilizes technology we are genuinely passionate about, like Artificial Intelligence and Cloud Computing.
Creating a product that can brighten up someone's day.

What we learned

Participating in UofTHacks reminded us that, above all, hackathons should be fun – it's what keeps the excitement alive for our projects.

What's next for ImagiTalk

Implementing chat history and contextual memory to facilitate a more nuanced chatting experience.
"Add character" feature that would allow the users to create a chat with the character of their choice.