Parallel Tracks

Have you ever wondered

Comment

Inspiration

We where inspired to create a tool that brings people together to discuss complex ethical issues in a fun, engaging, and accessible way. We believe that technology can be a powerful force for good, and we wanted to build something that could be used in to foster empathy and critical thinking.

What it does

ParallelTracks is a real-time, interactive platform for exploring ethical qualms.

It gamifies the classic "trolley problem" for a live audience, transforming it into an engaging experience.

Here's how it works:

  • For the Host: A community leader or just a member of a group of friends can present a moral scenario to the group. Our platform then uses generative AI to bring the scenario to life, creating a unique, cartoon-style image of the dilemma and generating a variety of AI-generated perspectives on the problem .
  • For the Audience: Participants use their mobile devices to cast their votes in real-time. To make the experience more inclusive and accessible, we've integrated ElevenLabs to provide high-quality, text-to-speech audio for every AI-generated response, ensuring that people with visual impairments or different learning styles can fully participate. A live leaderboard tracks the group's collective decisions, providing a starting point for a rich, data-driven discussion.

##How we built it

  • Frontend (Voting App): The user-facing voting application is a Next.js and React app built with TypeScript and styled with Tailwind CSS.

  • Backend (Host Display & AI Orchestration): The host's display and our AI orchestration layer are powered by a Flask and Python server.

  • Database & Real-Time: We use Supabase for our database and to manage the real-time voting and leaderboard updates.

  • Generative AI:

    • OpenRouter is our gateway to a suite of cutting-edge AI models. We use it to generate:
      • Images with Google's Gemini 3 Pro Image Preview.
      • Text-based ethical arguments from grok, gemini, chatGPT, claude, and deepseek.
    • ElevenLabs API is used for high-quality, low-latency text-to-speech, providing an essential layer of accessibility to our platform.

Challenges we ran into

Our biggest challenge was orchestrating multiple, independent AI services in real-time. Generating an image, five different text responses, and five corresponding audio files—all in a matter of seconds—required careful asynchronous management and robust error handling. We also spent a significant amount of time fine-tuning our prompts for the image generation model to ensure a consistent and appropriate visual style for our sensitive subject matter.

Accomplishments that we're proud of

We are incredibly proud of creating a fully functional, multi-modal, and real-time application that is not just a technical showcase, but a tool with real social value. Integrating text, image, and audio generation into a seamless user experience was a major accomplishment. Most importantly, we're proud to have built a "Community Class Act" that has the potential to make difficult conversations more engaging, inclusive, and productive.

What we learned

This project taught us the immense power of combining different AI models to create a rich, multi-sensory experience. We learned how to build a real-time, full-stack application from the ground up, and we gained a deeper appreciation for the importance of accessibility in web design. Above all, we learned that technology can be a powerful tool for bringing communities together and fostering the kind of thoughtful dialogue that is more important now than ever.

Built With

Share this project:

Updates