Inspiration
I made Polar Search because I love exploring creepy, weird, or forgotten corners of the internet like conspiracies, urban legends and lost media so I thought this was an ideal use case for the Sonar model. Inspired by the iceberg meme format, the app structures information into an interactive tierlist, where each tier reveals a deeper, weirder, or more specialized insight — letting users explore subjects like internet mysteries, fringe science, etc.
What it does
Polar Search is a knowledge exploration tool that organizes search results into an interactive iceberg visualization. Each query generates five tiers of information depth — from surface-level facts to obscure, long-tail insights.
Users can:
Click any subtopic to trigger deep research using sonar-reasoning-pro
Generate Research Capsules with curated sources
Build Timelines to trace the history or evolution of a topic
Technologies
Frontend: Built in React.js with MUI for responsive UI and interactive iceberg navigation
Backend: Python with FastAPI, handling all Sonar API calls and result shaping
Perplexity Sonar APIs: Used sonar for topic clustering and sonar-reasoning-pro for advanced research, content retrieval, and timeline synthesis
Challenges
Structuring Sonar’s output into a clear, tiered format meant I had to experiment with prompt design to make it visually meaningful whilst also achieving the desired research output and format
Making the iceberg UI both interactive and responsive was difficult as I had to balance the functionality with aesthetics
Managing API rate limits while enabling multi-step queries (e.g., follow-ups and timeline generation)
What's next for Polar Search
- Collaborative features - Users can build and share their own icebergs with notes and context
- ** Saveable Icebergs** - Letting users bookmark and revisit their knowledge maps
Log in or sign up for Devpost to join the conversation.