Skip to content
 
 

Repository files navigation

🤖 MeetSync: Multi-Agent Meeting Intelligence Ecosystem

MeetSync is not just a video calling platform—it's a sophisticated Multi-Agent AI Ecosystem designed to automate meeting management, commitment tracking, and enterprise-grade accountability. By deploying a suite of specialized AI Agents, MeetSync transforms every conversation into a structured database of actionable intelligence.

MeetSync Banner Accountability Engine Groq Powered Next.js Firebase Tailwind CSS


🧠 The Multi-Agent Intelligence System

MeetSync operates using three core specialized agents that process meeting data in a pipeline to ensure maximum accuracy and depth.

1. 📋 The Intelligence Extraction Agent

This agent performs the primary post-meeting analysis, converting audio into high-fidelity structured data.

  • Deep Summary & Sentiment: Goes beyond basic notes to capture the emotional tone and outcomes.
  • Commitment Intelligence: Automatically identifies promises, obligations, and deliverables.
  • Risk & Blocker Detection: Flags potential future problems (risks) versus active progress-preventers (blockers).
  • Structured Output: Generates valid JSON intelligence covering decisions, questions raised, and participant intent.

2. ⚖️ The Accountability Engine (Agent 3)

The "Brain" of MeetSync's persistence. It compares the current meeting with your organization's entire history.

  • Historical Alignment: Syncs current promises with past commitments to check for consistency.
  • Reliability Scoring: Assigns a Reliability Score (0-100) to participants based on their track record of completing deliverables.
  • Pattern Recognition: Detects repeated delays or circular commitments (e.g., the same approval promised in three consecutive meetings).
  • Escalation Alerts: Automatically identifies if an overdue task creates a critical downstream launch risk.

3. 🎯 The Strategic Preparation Agent

Acts as your personal advisor before you even enter the room.

  • Executive Briefings: Summarizes relationship history and project status in 3-5 high-impact sentences.
  • Recommended Strategy: Advises on how to approach a meeting based on the participants' historical reliability and open items.
  • Dynamic Questioning: Generates "hard questions" to move projects forward based on identified blockers.

🚀 Advanced Analysis Features

  • Interactive Meeting Q&A: Each recording becomes a searchable knowledge base. Ask "Who agreed to the API change?" or "What was the budget concern?" and get cited answers instantly.
  • Relationship Intelligence: Understand participant dynamics over time, including consistency, sentiment trends, and past collaboration hurdles.
  • Real-time Intelligence Pipeline: Processes video recordings directly from Stream.io through Groq's Whisper-large-v3 and Llama-3.3-70b for near-instant results.

🛠️ The Technology Behind the Intelligence

Layer Component Implementation
LLM Inference Groq LPU™ Blazing-fast Llama 3.3 70B & Whisper-v3
Memory Cloud Firestore Persistent storage for historical commitments & reliability matrices
Video Engine Stream.io High-fidelity recording and low-latency conferencing
Framework Next.js 14 Secure Server Actions and highly responsive UI

🎨 Aesthetic: Parchment Modern

MeetSync utilizes a custom Parchment Modern design system. This isn't just a theme; it's a productivity-focused interface that combines high-contrast typography (Playfair Display) with warm, non-fatiguing backgrounds, making long analysis sessions comfortable for the eyes.


📦 Getting Started

Prerequisites

  • Node.js 18+
  • Groq API Key (Llama 3.3 and Whisper support)
  • Firebase (Auth & Firestore)
  • Stream.io (Video SDK)

Environment Setup

# AI & INFRASTRUCTURE
GROQ_API_KEY=gsk_...
NEXT_PUBLIC_STREAM_API_KEY=...
STREAM_SECRET_KEY=...

# PERSISTENCE
NEXT_PUBLIC_FIREBASE_API_KEY=...
# ... other firebase keys

Quick Run

npm install
npm run dev

📄 License

Distributed under the MIT License.


MeetSyncFrom Conversations to Commitment Intelligence.

About

An AI executive assistant that reads meeting notes, stores important memories, learns from every meeting, and automatically prepares users for future meetings using everything it remembers.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages