Serverless Video Analyzer

A full-stack serverless application for automated technical video analysis using AWS (Lambda, S3, DynamoDB, API Gateway) and OpenAI (GPT-4o).

Features

• Upload: Drag & drop video uploads directly to S3 via presigned URLs.
• Analysis: Automated extraction of keyframes and analysis of optical properties (lighting, focus, composition).
• Context-Aware: Uses chat history to guide the AI's analysis.
• Serverless: Zero server management. Scales automatically.
• Infrastructure as Code: Fully provisioned via Terraform.

Architecture

1. Frontend: Next.js (App Router, Tailwind CSS, Shadcn UI).
2. API: Next.js API Routes + AWS API Gateway (S3 Proxy).
3. Storage:
   • video-analyzer-videos-* (Raw uploads)
   • video-analyzer-results-* (JSON Analysis)
   • DynamoDB InterviewAnalysis (Status tracking)
4. Compute:
   • Worker Lambda: Python (OpenCV + OpenAI SDK). Triggered by S3 upload.

Getting Started

Prerequisites

• Node.js 18+
• Python 3.9+
• AWS CLI configured
• Terraform installed

Installation

1. Clone the repo:

   git clone https://github.com/Ris345/Serverless-Video-Analyzer.git
   cd Serverless-Video-Analyzer

2. Install dependencies:

   npm install
   pip install -r requirements.txt

Local Development

npm run dev

Open http://localhost:3000.

Deployment

1. Infrastructure (Terraform)

cd terraform
terraform init
terraform apply

2. Backend (Lambda)

./deploy_worker.sh

Security

• Secrets: Managed via .env (not committed).
• IAM: Least-privilege roles for Lambda and API Gateway.
• Access: Presigned URLs for secure S3 uploads.