About Vega
Revolutionary AI-powered video analysis platform that simulates real human viewers to provide unprecedented insights into video content engagement and effectiveness.
How Vega Works
1. Video Upload
Upload your video content through our secure, authenticated interface
2. AI Agent Analysis
24 specialized AI agents simulate diverse viewer personas and analyze content simultaneously
3. Comprehensive Report
Receive detailed insights on engagement, effectiveness, and audience response patterns
Frontend Architecture
- Next.js 15 with App Router & Turbopack
- React 19 with modern hooks & concurrent features
- Tailwind CSS for responsive, cosmic-themed UI
- Framer Motion for smooth animations & transitions
- Supabase Auth with OAuth integration
Backend & AI Engine
- Google ADK 1.15.1 for AI agent orchestration
- Gemini 2.0 Flash for lightning-fast video processing
- Multi-Agent System with 24 parallel viewer personas
- FastAPI backend with async processing
- FFmpeg integration for video processing
Intelligent Multi-Agent System
Agent Pipeline
- 1Video Transcriber: Complete audio-to-text conversion with speaker identification
- 2Transcript Summarizer: Key content extraction and context analysis
- 324 Reviewer Agents: Parallel analysis from diverse viewer perspectives
- 4Synthesis Agent: Comprehensive report generation and insights
Viewer Personas
Demographics: Age groups, cultural backgrounds, education levels
Interests: Tech enthusiasts, casual viewers, subject matter experts
Viewing Styles: Critical analyzers, entertainment seekers, learners
Engagement Patterns: Active participants, passive consumers, social sharers
Lightning Fast
Powered by Gemini for rapid video processing and real-time insights
Multi-Perspective Analysis
25+ AI agents simulate diverse viewer demographics and engagement patterns
Deep Insights
Advanced AI understanding of context, emotion, and audience engagement
Built for ShellHacks 2025
Explore our codebase or fork it to create your own AI-powered video analysis platform.