Trace

A transparency tool for analyzing conflicts of interest in research, journalism, and public discourse

Comment

Inspiration

E.O Wilson said, we're drowning in information but starving for wisdom. Turns out, we're also starving for receipts.

The problem with the wild spread of misinformation isn't just information overload, it's that we can't tell who's behind what we're reading. Social media has democratized voices but also disguised agendas: influencers pose as experts, sponsored content masquerades as journalism, and paid experts present opinions as facts. Our goal with Trace is to cut through that noise by revealing the conflicts of interest hiding behind the information we consume. We believe that transparency isn't just nice to have, but the foundation of critical thinking.

What it does

Trace analyzes articles, research papers, and public statements to reveal:

  • Funding sources
  • Institutional affiliations of authors and experts
  • Financial conflicts of interest that might bias information
  • Connection maps showing relationships between sources, funders, and stakeholders

Users can paste a URL, and Trace generates a transparency report showing who benefits from the narrative being presented.

How we built it

  • Frontend: React + TypeScript for a responsive, type-safe interface
  • Design: Figma for prototyping and UI/UX design
  • Backend: FastAPI (Python) for high-performance API endpoints
  • AI: Gemini 3 for intelligent extraction of conflicts, entities, and relationships from unstructured text
  • Data: SerpAPI for real-time web searches to cross-reference authors, institutions, and funding sources
  • Workflow: User submits URL → SerpAPI gathers context → Gemini analyzes conflicts → FastAPI serves results → React displays transparency insights

Challenges we ran into

  • Real-time processing: Orchestrating multiple API calls (Serp (Youtube Transcript API) + Gemini)
  • Time Constraint

Accomplishments that we're proud of

  • Built a working full-stack prototype in 24 hours that actually surfaces hidden conflicts
  • Successfully chained SerpAPI + Gemini to create an intelligent research pipeline
  • Created an intuitive React interface that makes complex relationships understandable at a glance
  • Designed a clean, professional UI in Figma and implemented it with TypeScript for reliability

What we learned

  • How fragmented and poorly standardized conflict of interest reporting really is
  • The power of combining search APIs with LLMs for research automation
  • Prompt engineering is as important as the model itself when extracting structured insights

What's next for Trace

  • Browser extension: Analyze articles as you browse with one click
  • Expand data sources: Include lobbying records, political donations, board memberships via additional APIs
  • Community verification: Let users submit and verify conflict information
  • Confidence scores: Show reliability metrics for each detected conflict
  • Export reports: Generate shareable PDF transparency reports for citing in research

Built With

Share this project:

Updates