Problem ~ solution

Our demo video was recorded with loom you can watch it here

demo video <<

Inspiration

I wanted to create a smarter way for teams to collaborate, where AI doesn’t just listen but understands, retrieves, and acts. Most video or chat tools only connect people; ConvoGenius connects people and knowledge. I was inspired by the idea of combining search, chat, and meetings into one AI-driven workspace powered by You.com’s APIs.

What it does

ConvoGenius is the AI-powered collaboration platform that integrates video conferencing, real-time chat, and AI-enhanced insights. It uses You.com Search APIs for contextual retrieval during live meetings and allowing users to ask natural-language questions and get verified, summarized answers instantly. The platform also includes:

Real-time transcription and summarization
Context-aware AI responses
Analytics dashboard for meetings
Secure authentication and user management

How we built it

Frontend: Built with Next.js and TypeScript for a fast, scalable UI.
Backend: Powered by tRPC for type-safe API communication.
Auth: Managed with Better Auth for secure sign-in.
Video & Chat: Integrated Stream SDKs for real-time video and messaging.
AI Engine: OpenAI GPT-4 for summarization and transcription intelligence.
Knowledge Retrieval: You.com APIs (Search, Images, and Chat) for RAG-based data augmentation.
Database: PostgreSQL (Neon) for persistence and structured storage.
Other Tools: Inngest for webhooks and event automation.

API usage is verifiable through You.com request logs

Challenges we ran into

Handling real-time synchronization between chat, video, and AI summaries.
Optimizing API response latency from You.com for smoother RAG integration.
Managing state consistency across multiple user sessions during live collaboration.
Designing a clean yet intuitive UI for both desktop and mobile users.

Accomplishments that we're proud of

Successfully integrated three You.com API endpoints for live retrieval and context augmentation.
Built a fully functional multi-user meeting system with transcription and AI chat.
Created a seamless user experience combining video, chat, and search.
Developed analytics tracking and event logging with Inngest and You.com logs.

What we learned

How to design RAG-based systems that blend search APIs with generative AI.
The importance of API observability and logging during production.
That great AI experiences require both human context and machine intelligence.

What's next for ConvoGenius

Expand to include AI-generated meeting notes synced to Notion or Google Docs.
Build browser extensions that allow quick access to ConvoGenius from any page.
Add voice command control using real-time transcription triggers.
Optimize You.com query caching for faster contextual results.