Picasso AI

Inspiration

Similarly to how tools like Cursor AI and Claude empower software developers, we wanted to build a platform that does the same for creative professionals. As artists ourselves, we've felt firsthand the friction in our workflows and have heard the same from peers time and time again. While countless productivity tools exist for coders, artists and designers still lack a streamlined, AI-powered assistant that enhances their creative process in real-time. That’s how Picasso AI was born: an intelligent, cross-platform creative companion built by artists, for artists.

What it does

Picasso AI is an all-in-one creative assistant for professionals in visual arts, design, and animation. It integrates OpenAI's GPT-4o for natural language interaction and concept ideation, while our custom-built interpolation engine allows for smooth frame generation between two images. Whether you're storyboarding, animating, or experimenting with style transitions, Picasso AI helps bring your artistic vision to life faster and with less friction.

How we built it

We built Picasso AI using the Electron framework, which allowed us to create a cross-platform desktop application with a sleek and responsive UI. At its core, the app uses GPT-4o for prompt understanding and creative assistance, while our own interpolation algorithm (heavily inspired by models like RIFE and DAIN) enables smooth image transitions. Due to hardware limitations, we couldn’t use existing GPU-heavy models directly, so we innovated a lightweight solution that approximates the behavior of those state-of-the-art interpolators using more accessible resources.

Challenges we ran into

One of our biggest challenges was the lack of resources and funding. Many of the models we initially considered required powerful GPUs or vast datasets, which were out of reach for our team. This constraint forced us to get creative and design our own interpolation method from the ground up. Balancing performance and accessibility was difficult, but it led to some of our most rewarding technical breakthroughs.

Accomplishments that we're proud of

We're proud that we were able to build a functional and user-friendly tool despite our hardware and resource limitations. Creating an interpolation algorithm that mimics state-of-the-art behavior without relying on massive infrastructure was a huge achievement. We're also proud of how seamless the user experience feels, thanks to thoughtful design and tight integration of GPT-4o’s capabilities.

What we learned

We learned that limitations can actually fuel creativity. Building our own interpolation method taught us a lot about motion estimation, temporal consistency, and optimization. We also deepened our understanding of AI integration within desktop applications and how to balance usability with performance in constrained environments.

What's next for Picasso AI

We hope to deploy Picasso AI into the real-world market to support creatives in animation, design, and visual storytelling. Our next steps include improving the interpolation quality, integrating additional AI-powered tools like inpainting and style transfer, and potentially offering cloud-based features for more intensive tasks. With proper funding and user feedback, Picasso AI has the potential to become the go-to assistant for artists everywhere.