Inspiration
Our motivation for creating "Off the Record" came from recognizing how often people withhold their ideas due to time constraints or self-censorship. We wanted to create a safe space where everyone feels comfortable sharing their thoughts openly via speech, enabling a rich exchange of ideas that might otherwise remain unexpressed.
What it does
"Off the Record" is a platform that facilitates anonymous knowledge sharing. Users engage in conversations with AI agents, which collect and analyze these discussions. The platform then generates a custom report that consolidates diverse perspectives, giving users insights into what others are thinking without revealing individual identities.
How we built it
Frontend - Next and Vercel Backend - Node + Python LLM - Anthropic and ChatGPT Data Source: Reddit (MindsDB facilitated) AgentOps - AgentOps
Challenges we ran into
We had a bunch of formatting issues in the LLM responses. Many of the functions were interdependent so we had to debug them carefully. We need to iterate to find the right set of instructions to pass to the AI to conduct the interview
Accomplishments that we're proud of
We are proud of creating a product that not only encourages the free exchange of ideas but does so in a way that respects user privacy. We also created a custom report for each participant which was neat.
What we learned
The agents are effective sound boards that can sharpen one's idea. We didn't expect to observe this. The AI agents aren't mere information gatherers but active participants in idea refinement.
What's next for Off the Record
Being able to do more structured data extraction agenetically to enable quantitative analysis on the text corpus.
Built With
- anthropic
- fastapi
- javascript
- openai
- python
- replit
- tailwind
- typescript
- vercel
Log in or sign up for Devpost to join the conversation.