Socratic AI

Inspiration

We were tired of news feeling like a tug of war where every outlet pulls in a different direction. People want clarity, not chaos. Socratic AI was born from the idea that balanced understanding should be accessible to everyone.

What it does

Socratic AI takes any topic and returns four perspectives: left, center, right, and a distilled fact-only view. It helps users quickly understand an issue without needing to dig through dozens of articles.

How we built it

We used React for the frontend, Flask as the secure API layer, and n8n as the backend workflow engine. n8n fetches articles from curated sources, summarizes them with LLMs, extracts neutral facts, and returns structured JSON to Flask, which sends it to the frontend. The entire system is built for fast iteration and smooth integration.

Challenges we ran into

Coordinating consistent JSON contracts between all layers was tough. Keeping API keys private, normalizing messy article formats, and getting LLMs to produce stable, structured summaries were challenges as well. Balancing accuracy and speed without breaking the multi-perspective vision took work.

Accomplishments that we're proud of

We created a clean, interpretable multi-service pipeline that actually works end to end. The four-lane perspective layout makes complex topics easier to digest, and we’re proud of making something that reduces confusion instead of adding to it.

What we learned

We learned the value of tight data contracts, how to make multiple systems cooperate under pressure, and how to guide LLMs into producing consistent, structured outputs. We also learned how to simplify complex news ecosystems into something digestible.

What's next for Socratic AI

More sources, smarter topic clustering, credibility scoring, user accounts, and deeper topic exploration. The long-term vision is to make Socratic AI the fastest and clearest way to understand any issue without bias getting in the way.