Skip to content

omkar42785/Adforge

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

45 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Samsung Ad-Forge


Submission for the Samsung PRISM Generative AI Hackathon 2025 - Theme 2: Content Creation

Team Name: Temperature=0.0


1. The Vision

Samsung Ad-Forge is a B2B generative AI platform that empowers Samsung's vast network of third-party retailers to create professional, hyper-local, and brand-safe video advertisements in seconds. Our platform transforms every retail promoter into a local marketing expert, closing the critical "last-mile" marketing gap and driving sales where they matter most.

2. Key Features

  • AI Creative Director: Leverages a powerful LLM to analyze a retailer's simple business goal and generate a complete, multi-scene creative blueprint for a 15-second ad.
  • On-the-Fly Video Generation: Utilizes a state-of-the-art text-to-video model to generate unique, cinematic visuals for the ad's core narrative.
  • Brand-Safe Hybrid Model: Intelligently fuses AI-generated scenes with curated, brand-perfect product shots to ensure 100% brand compliance.
  • Dynamic AI Voiceover & Music: Generates a unique voiceover and a fitting background music track for every ad, creating a fully autonomous creative pipeline.
  • Professional Asynchronous Architecture: Built on a robust, scalable, serverless backend that handles long-running generative tasks efficiently, providing a seamless user experience.

3. System Architecture

Ad-Forge System Architecture

Our system is built on a modern, asynchronous, serverless architecture using AWS. The workflow is orchestrated by three distinct microservices:

  • A `/forge endpoint to rapidly accept and queue new ad generation jobs.

  • A long-running worker function that executes the entire multi-stage generative AI pipeline.

  • A `/status endpoint for the client to poll for the final, completed video URL.

4. Tech Stack

  • Frontend: Flutter
  • Backend: AWS Lambda (Python), API Gateway, DynamoDB, S3
  • AI Services:
    • LLM (Blueprint): OpenRouter (Google Gemini Flash)
    • Text-to-Video: Hugging Face Inference Client (Fal.ai provider)
    • Text-to-Speech: ElevenLabs
    • Video Assembly: Shotstack

5. How to Run the Project

Prerequisites

  • AWS CLI & SAM CLI
  • Docker
  • Flutter SDK
  • An AWS account with configured credentials
  • API keys for the services listed above, stored in AWS Systems Manager Parameter Store.

Backend Deployment

# Navigate to the backend directory
cd backend/
# Build the serverless application
sam build
# Deploy the stack to your AWS account
sam deploy --guided

Frontend Setup

# Navigate to the frontend directory
cd flutter-app/
# Install dependencies
flutter pub get
# Run the app
flutter run

other Ads generated by AdForge

WhatsApp.Video.2025-08-29.at.00.01.00.mp4
5e64b6c5-e4d7-40d5-8434-80bde93c7105.mp4
2240477e-6d0c-4595-bc4f-165a064fde25.mp4

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors