ArtGems

• 

  ARTGEMS interface

Inspiration

We’ve seen how often the same names dominate art history and curatorial narratives-leaving out diverse voices and underrepresented artists. I've built ArtGems to make hidden artistic treasures accessible to curators, educators, and researchers using the power of AI. The goal: democratize access to global art history and promote cultural diversity in curatorial research.

What it does

ArtGems is an AI-powered art discovery assistant. Users input curatorial prompts (e.g., "artists exploring climate change"), and the system returns a curated list of underrepresented artists relevant to that theme-complete with biographies, artistic styles, notable works, and verified sources. The insights are fetched dynamically using the Perplexity Sonar Deep Research API, ensuring each result is real-time, cited, and reliable.

How I built it

Frontend: HTML/CSS and Flask templates Backend: Flask app using Python and Perplexity’s Sonar Deep Research API API Integration: User queries are processed in natural language, send them to Sonar, extract key insights and citations, and display them in a structured format Deployment: Local Flask server for demonstration and testing

Challenges I ran into

Bias and Representation: Balancing accuracy with the goal of surfacing lesser-known artists was challenging-mainstream names often dominated early responses.

Accomplishments that I am proud of

I built a working prototype in less than 6 hours with real-time, dynamic insights Integrated a complex research API and made it accessible to non-technical users Delivered real, curated outputs highlighting artists from underrepresented backgrounds Maintained a clean and usable UX

What I learned

How to design inclusive AI systems for cultural use cases The importance of system prompt tuning in controlling bias and optimizing relevance API integration challenges in generative research tools

What's next for ArtGems

Add filters for medium, region, time period, and budget constraints Enable follow-up queries and conversational refinement Partner with museums, curators, and educators for real-world use Build a dataset of verified, diverse artists to complement API searches Deploy to the web and mobile for wider access

Built With

• flask
• perplexity
• python
• sonar