Refyna

• 

  This is the home/Dashboard

• 

  This is after the AI has annotated.

• 

  This is the design session history

• 

  This is when the AI is thinking

• 

  This is the settings screen

• 

  These are the modules for the design dojo

• 

  This is when the quiz is about to begin

• 

  This is an example of the quiz

• 

  This is an example of when the Ai corrects you

• 

  More images about how the leaderboard looks like

• 

  This is an example of how the leaderboard

• 

  This is an example of annotated feedback

Inspiration

Design is evolving rapidly, especially with the rise of AI. While tools are becoming faster and more powerful, we noticed a growing gap: design fundamentals are being skipped. Many designers can execute quickly, but struggle to explain why a design works, make confident decisions, or grow beyond a certain level.

We were inspired to build Refyna AI to address this gap. We wanted to create a platform that doesn’t replace designers, but refines them by helping designers strengthen their fundamentals, build better judgment, and stay relevant in a future where intentional design will matter more than ever.

What it does

Refyna AI is an intelligent design growth platform that helps designers improve through AI-powered feedback and deliberate practice. It has two core modules:

Design Audit: Designers can upload or paste their designs (including Frameshare tokens), and Refyna AI provides annotated feedback highlighting areas that need improvement, with clear explanations. Designers can also converse with the AI to dive deeper and download audit reports.

Design Dojo: A practice environment where designers sharpen their fundamentals—typography, color, layout, hierarchy, accessibility, through interactive quizzes and fast feedback.

Together, these modules help designers practice, reflect, and grow consistently.

How we built it

We built Refyna AI by combining: A design-trained AI feedback system Visual annotation logic for targeted critique A modular architecture to support multiple learning experiences A quiz engine designed around fast feedback and engagement We also tightly integrated Refyna AI into the broader Frameverse ecosystem, allowing seamless workflows like reviewing real design assets shared via Frameshare. Throughout development, we worked iteratively, testing internally, releasing early demos, and gathering direct feedback from designers in our community.

Challenges we ran into

One of the biggest challenges was ensuring the AI feedback felt useful, contextual, and encouraging, not generic or judgmental. Design feedback is nuanced, and getting the tone right was critical.

Another challenge was balancing ambition with focus. We had many ideas; from leaderboards to learning paths; but had to prioritize building a strong core experience before expanding.

We also learned that designers care deeply about clarity and trust, so every interaction had to feel intentional.

Accomplishments that we're proud of

Successfully built an AI system that provides annotated, design-aware feedback Created a practice module that makes learning fundamentals feel engaging, not academic Launched an early demo and onboarded real designers to test Refyna AI Built an active community that gives honest, thoughtful feedback Positioned Refyna AI as part of a larger ecosystem, not a standalone tool

What we learned

We learned that:

Designers want feedback, but only if it’s respectful and actionable Fundamentals still matter deeply, regardless of experience level Community-driven development leads to better product decisions AI is most powerful when it augments human thinking, not replaces it Most importantly, we learned that designers are eager for tools that help them grow long-term, not just ship faster.

What's next for Refyna

Next, we plan to: Complete the Design Dojo experience with leaderboards, UXD points, and learning recommendations Personalize learning paths based on user weaknesses Improve feedback accuracy and performance Expand Refyna AI into educational institutions and design communities Connect skill progression to real opportunities within the Frameverse ecosystem Our long-term vision is for Refyna AI to become the place where designers continuously refine their craft and turn mastery into opportunity.

Built With

• gemini
• react
• render
• supabase
• typescript