SyncUp - AI Meeting Assistant for Google Meet

A Chrome extension that captures conversations in Google Meet and automatically generates contextual information cards using AI. Built with Gemini API for transcription and Cerebras with Meta Llama/gpt-oss-120b for Fast and intelligent content generation.

Resources

• Demo video: https://youtu.be/NgmFNfwHOZI
• Landing Page and installation: https://rohit026.github.io/
• Presentation link: https://drive.google.com/file/d/1bPaV5dkZIV4J0YWZTHp43G4R7cHjVM-i/view?usp=sharing
• Installation link (Chrome extension in review by the Google team): https://drive.google.com/drive/folders/10416hfb4tRKtLIdB_TvXWAOM84pDJBkm?usp=sharing

What It Does

SyncUp listens to your Google Meet conversations and automatically creates information cards about topics mentioned. When someone talks about Docker, Kubernetes, or any technical concept, the extension generates a detailed explanation card with key points, use cases, and resources - like having ChatGPT instantly explain things during your meeting.

The extension also includes an AI chatbox where you can ask questions about the meeting or any topic, and get context-aware responses.

Key Features

Real-Time Topic Detection

• Captures speech from all meeting participants using Google Meet captions
• Extracts important keywords and technical terms from conversations
• Generates detailed explanation cards automatically

AI-Powered Information Cards

• Each card includes a summary, key points, use cases, and learning resources
• Cards expand/collapse for easy reading
• Prevents duplicate cards for the same topic

AI Chatbox

• Ask questions about the meeting or general topics
• Get instant AI responses with meeting context
• Works like ChatGPT but aware of your meeting content

Multilingual Support

• Supports English and Hindi conversations
• Detects code-switching between languages
• Responds in English regardless of input language

Clean Interface

• Sidebar that slides in from the right side of Google Meet
• Toggle button to show/hide the sidebar
• Start/Stop/Clear controls
• Smooth animations and modern design

Upcoming Features (1 month)

• Meeting Memory for post-meeting analysis, questions, and cross-meeting reference
• Multi-format attachments like PDFs, JPGs, MP4s, PPTs, etc.
• Enhanced UI with new themes and trending skins for Google meetings.
• Multi-platform support for Zoom, Microsoft teams, etc.

Technology Stack

• Chrome Extension (Manifest V3)
• Gemini API for transcription
• Cerebras API with Meta Llama 3.1-8B/gpt-oss-120b for content generation
• Web Speech API for local speech capture
• Google Meet caption integration

Installation

1. Clone this repository

git clone https://github.com/awneet23/Syncup.git
cd Syncup

2. Add your API keys in background.js (around line 18-19):

this.GEMINI_API_KEY = 'your_gemini_api_key_here';
this.CEREBRAS_API_KEY = 'your_cerebras_api_key_here';

Get API keys:

• Gemini API: https://ai.google.dev/
• Cerebras API: https://cerebras.ai/

3. Load the extension in Chrome:

• Open chrome://extensions/
• Enable "Developer mode" (top right toggle)
• Click "Load unpacked"
• Select the syncup folder

4. Pin the extension to your toolbar for easy access

How to Use

1. Join a Google Meet call
2. Click the toggle button in the middle-right corner to open the sidebar
3. Click the "Start" button (play icon) in the sidebar
4. Enable Google Meet captions by pressing the "C" key - this captures all participants
5. Start talking or let others speak
6. Watch as information cards appear automatically for mentioned topics
7. Click "Read more" on any card to see full details
8. Use the chatbox at the bottom to ask questions

The extension processes speech in 15-second batches and generates cards for important keywords it detects.

Example Usage

Scenario 1: Technical Discussion

• Someone mentions "Docker containerization"
• Extension detects "Docker" as a keyword
• Generates a card explaining Docker with key points and resources
• Card appears in the sidebar within 2-3 seconds

Scenario 2: Using the Chatbox

• You type "What is Kubernetes?"
• AI responds with a comprehensive explanation
• If Kubernetes was mentioned in the meeting, response includes that context

Scenario 3: Multilingual Meeting

• Participant says "हमें Docker install करना है" (We need to install Docker)
• Extension detects Hindi, extracts "Docker"
• Generates card in English with Docker information

Architecture

The extension has three main components:

Content Script (content_script.js)

• Injects the sidebar UI into Google Meet pages
• Captures speech using Web Speech API and Google Meet captions
• Displays information cards and handles user interactions
• Manages the chatbox interface

Background Service Worker (background.js)

• Processes transcripts and extracts keywords using Cerebras API
• Generates detailed information cards using AI
• Handles chatbox questions with context awareness
• Manages communication between components

Popup (popup.html, popup.js)

• Extension control panel
• Shows recording status and session stats
• Provides Start/Stop/Clear controls

API Integration

Cerebras API with Meta Llama/gpt-oss-120b

• Model: gpt-oss-120b / Llama 3.1-8B
• Used for keyword extraction from transcripts
• Generates comprehensive explanations for detected topics
• Powers the chatbox Q&A functionality

Gemini API

• Configured for future transcription enhancements
• Currently using Web Speech API and Meet captions

Performance

• Keyword detection: Real-time (less than 1 second)
• Card generation: 2-3 seconds per topic
• Chatbox response: 1-2 seconds
• Memory usage: Under 100 MB
• Minimal impact on meeting performance

Hackathon Highlights

Sponsor Technology Integration

• Gemini API: Configured for AI processing
• Cerebras API: Core inference engine using Llama 3.1-8B/gpt-oss-120b
• Meta Llama/gpt-oss-120b: Powers all content generation
• Docker: Containerization ready

Innovation

• Captures all participants, not just the user
• Real-time processing during meetings
• Context-aware AI chatbox
• Multilingual support with code-switching
• Non-intrusive sidebar design

Problem Solved Traditional meetings require manual note-taking and context switching to look up unfamiliar terms. SyncUp automatically generates explanations for mentioned topics in real-time, letting participants stay focused on the conversation while still learning about new concepts.

Troubleshooting

No cards appearing:

• Make sure you clicked "Start" in the sidebar
• Enable Google Meet captions by pressing "C"
• Check that API keys are configured in background.js
• Look for errors in the browser console (F12)

Sidebar not showing:

• Click the toggle button (magnifying glass) in bottom-right
• Refresh the Google Meet page
• Check that you're on meet.google.com

Cards not generating for topics:

• Verify Cerebras API key is valid
• Check browser console for API errors
• Make sure topics are clearly mentioned in conversation
• Wait 15 seconds for batch processing

Development

Project Structure

syncup/
├── manifest.json          # Extension configuration
├── background.js          # Service worker (AI processing)
├── content_script.js      # Sidebar UI and speech capture
├── popup.html/js          # Extension popup
├── styles.css             # Sidebar styling
└── icons/                 # Extension icons

Key Functions

In background.js:

• extractTopicsAndGenerateCards() - Extracts keywords using Cerebras
• generateAIExplanationCard() - Creates detailed cards for topics
• handleChatboxQuestion() - Processes chatbox queries

In content_script.js:

• startMeetCaptionCapture() - Captures Google Meet captions
• startMicrophoneRecognition() - Captures user's speech
• updateContextualCards() - Renders cards in sidebar

Acknowledgments

Built for hackathon submission using:

• Google Gemini API
• Cerebras Cloud Platform
• Meta Llama 3.1-8B model
• gpt-oss-120b
• Chrome Extension APIs

---

Note: This extension requires valid API keys for Gemini and Cerebras to function. Demo mode is not currently active - you must configure your own API keys.