Inspiration
What it does
How we built it
Challenges we ran into## Inspiration...
Practicing competitive debate or preparing for public speaking can be difficult when you don't have a partner available to drill against. Inspired by the need for accessible, on-demand debate prep, I built an intelligent sparring platform that adapts to various debate formats, motions, and stances to simulate realistic rounds.
What it does
- Dynamic AI Opponents: Users can select debate resolutions, pick affirmative or negative stances, and spar against a responsive AI adversary that builds counterarguments on the fly.
- Structured Flow: Guides users through structured debate mechanics, helping them practice constructive speeches, rebuttals, and cross-examinations.
- Feedback & Scoring: Evaluates arguments and provides analytical feedback scores to help debaters identify logical gaps and improve their delivery.
How we built it / How I built it
- Development Environment: Coded efficiently using Cursor for rapid prototyping and code structuring.
- Frontend & UI: Built using Streamlit for a clean, fast, and interactive web interface.
- AI & Logic: Integrated the Google GenAI SDK powered by Gemini models to handle complex argument generation, strategy formulation, and dynamic speech analysis.
- Deployment: Hosted live on Streamlit Cloud with secure environment management via TOML-based secrets.
- Code & Version Control: Managed and structured through Git and GitHub for continuous deployment.
Challenges I ran into
Managing environment configurations and API keys securely during cloud deployment, and structuring prompt engineering so the AI maintains a strict, objective, and competitive debate tone rather than acting like a standard chatbot.
What I learned
Advanced integration of generative AI APIs into real-time web UI components, and streamlining the DevOps workflow from local testing to continuous cloud delivery using GitHub and Streamlit Cloud.
Log in or sign up for Devpost to join the conversation.