Meeting Mind

Inspiration

Remote and hybrid work has increased the number of meetings, making it easy to lose track of discussions, action items, and key decisions.

What it does

Meeting Mind records meeting audio and processes it to produce transcriptions, summaries, and action items. The tool makes it easier for users to review discussions and follow up on next steps.

How we built it

Transcription and Analysis: Utilized Google Cloud Speech-to-Text for accurate transcription and the Gemini API for summarization, task extraction, and sentiment analysis.

Processing: Implemented both real-time and batch processing to handle live meetings and uploaded recordings.

Full-stack Development: Built the web app using Supabase for authentication and storage, Node.js with Express for the back end, and HTML/CSS/JavaScript for the front end.

Security: Applied input sanitization, API security measures, and integrated a reverse proxy with NGINX along with tools like Suricata for IDS/IPS.

Challenges we ran into

Time Constraints: Completing a full-stack app with integrated AI services in 48 hours was demanding.

Real-time Transcription: Ensuring low latency and maintaining transcription accuracy with live audio streams was difficult.

Multi-API Coordination: Managing and formatting responses from multiple APIs (transcription, summarization, and analysis) required careful planning.

Security Integration: Configuring security tools like Suricata, iptables, and NGINX on our Kali Linux setup while routing traffic effectively posed additional complexity.