Ads Genie

AdsGenie: Where Every Creative Wish Becomes Reality Presentation Link: https://docs.google.com/presentation/d/1pQDTEVCR3tsyqkQ4l2X8iehVstCKPIzDATnM_xKcMt0/edit?usp=sharing

Comment

AdsGenie - AI-Powered Video Ad Generator

Inspiration

Creating compelling video advertisements has always been a complex, time-consuming, and expensive process. Ad creators face numerous challenges:

  • High Production Costs: Traditional video production requires expensive equipment, professional videographers, actors, and post-production teams, making it inaccessible for small businesses and startups.

  • Time-Intensive Workflows: From concept to final video, the process can take weeks or months, with multiple rounds of revisions and coordination between different teams.

  • Limited Creative Flexibility: Once a video is shot, making changes requires re-shooting entire scenes, leading to additional costs and delays.

  • Resource Constraints: Small businesses and individual creators often lack the technical expertise, budget, or resources to produce high-quality video ads that compete with larger brands.

  • Audio Production Complexity: Adding appropriate background music, voiceovers, and sound effects requires separate tools, licensing, and audio engineering skills.

AdsGenie was born from the vision that every creator, regardless of their budget or technical expertise, should be able to bring their wildest creative ideas to life. Whether it's a dramatic product launch, an emotional brand story, a fun social media snippet, or a professional corporate video—AdsGenie fulfills every wish, transforming imagination into stunning video advertisements that captivate audiences.

What it does

AdsGenie is an end-to-end AI-powered platform that transforms creative ideas into professional video advertisements in minutes, not weeks. The platform empowers creators to:

🎬 Complete Video Creation Workflow

  • Create projects with customizable aspect ratios (16:9 for YouTube, 9:16 for TikTok/Instagram)
  • Build multi-scene video timelines with drag-and-drop editing
  • Define global characters and settings that persist across scenes
  • Generate AI-powered images and videos for each scene based on natural language descriptions

🎨 Intelligent Content Generation

  • Visual Generation: Uses advanced AI (Google GenAI, VEO3) to generate high-quality images and videos from text descriptions
  • Character Consistency: Maintains consistent character appearance across multiple scenes
  • Setting Preservation: Ensures visual coherence with persistent background settings

🎵 Advanced Audio Integration

  • Smart Music Selection: AI analyzes video descriptions to automatically select appropriate background music (e.g., sad music for emotional scenes, upbeat music for celebrations)
  • Text-to-Speech: Converts script to natural-sounding voiceovers with customizable voices
  • Audio Mixing: Seamlessly combines background music and voiceovers with proper volume balancing

⚡ Real-Time Preview & Export

  • Preview complete video projects before final generation
  • Export high-quality videos ready for social media, websites, or advertising platforms
  • Save and manage multiple projects for future editing

How we built it

Frontend (Next.js 16 + React 19)

  • Modern, responsive web interface with real-time drag-and-drop scene reordering
  • Interactive timeline and seamless backend API integration

Backend (FastAPI + Python)

  • RESTful API architecture with asynchronous task processing for video/audio generation
  • Background job management and webhook system for real-time status updates

AI Services Integration

  • Google GenAI (NanoBanana/VEO3): For image and video generation from text prompts
  • OpenAI GPT-4: Intelligent music type determination based on scene descriptions
  • ElevenLabs API: Text-to-speech and background music generation
  • Smart prompt engineering to ensure high-quality, contextually appropriate outputs

Cloud Infrastructure (AWS)

  • DynamoDB: Fast, scalable database for project and scene metadata
  • S3: Reliable storage for generated images, videos, and audio files
  • Lightsail: Cost-effective hosting for the application stack
  • Secure credential management and environment-based configuration

Audio Processing

  • FFmpeg integration for audio mixing and video stitching
  • Volume balancing between background music and voiceovers

Challenges we ran into

1. Instance Resource Constraints

  • Initially deployed on a nano Lightsail instance (512 MB RAM), which was insufficient for Next.js builds and concurrent operations
  • Solution: Upgraded to small instance (2 GB RAM) and optimized build processes

2. Network and Firewall Configuration

  • Multiple layers of firewall rules (Lightsail, UFW) blocking necessary ports
  • Solution: Systematically opened ports 3000 (frontend) and 8000 (backend) across all firewall layers

3. CORS Configuration

  • Frontend requests blocked due to missing CORS origins for the production URL
  • Solution: Updated CORS middleware to include the deployed frontend URL

4. Audio Generation API Complexity

  • ElevenLabs API endpoint changes and documentation updates
  • Solution: Implemented fallback mechanisms between Music API and Sound Effects API

5. Deployment Automation

  • Creating comprehensive deployment scripts that handle instance creation, snapshot management, and service setup
  • Solution: Built modular scripts for each deployment phase with proper error handling

Accomplishments that we're proud of

🚀 Complete End-to-End Solution

  • Built a fully functional, production-ready application from scratch in a hackathon timeframe
  • Successfully integrated multiple AI services (Google GenAI, OpenAI, ElevenLabs) into a cohesive platform

🎯 Intelligent Audio Generation

  • Implemented AI-powered music selection that analyzes video descriptions and automatically chooses appropriate background music
  • Created seamless audio mixing pipeline combining TTS and background music

⚡ Performance & Scalability

  • Designed cloud-native architecture that can scale with demand
  • Implemented asynchronous processing for resource-intensive operations
  • Optimized for cost-effectiveness using AWS Lightsail

🎨 User Experience

  • Created an intuitive, drag-and-drop interface that makes video creation accessible to non-technical users
  • Implemented real-time preview and editing capabilities
  • Built responsive design that works across devices

📦 Deployment Automation

  • Created comprehensive deployment scripts that automate the entire infrastructure setup
  • Successfully deployed to AWS with all services running smoothly

What we learned

Technical Learnings

  • Cloud Architecture: Gained deep understanding of AWS services (Lightsail, DynamoDB, S3) and how to integrate them effectively
  • AI Service Integration: Learned to work with multiple AI APIs, handle rate limits, and implement fallback strategies
  • Frontend-Backend Communication: Mastered CORS configuration, webhook implementation, and real-time status updates
  • Deployment Automation: Developed expertise in creating robust deployment scripts and managing infrastructure as code

Problem-Solving Skills

  • Learned to debug complex issues across multiple layers (network, application, cloud services)
  • Developed systematic approaches to troubleshooting deployment and configuration issues
  • Gained experience in optimizing resource usage and cost management

AI & Machine Learning

  • Discovered the power of prompt engineering for consistent, high-quality AI outputs
  • Learned to combine multiple AI services to create more intelligent solutions
  • Understood the importance of context-aware AI decision-making (e.g., music selection)

What's next for Ads Genie

🎬 Enhanced Video Generation

  • Support for longer video formats (30 seconds, 60 seconds, full commercials)
  • Advanced scene transitions and effects
  • Multiple video export formats and resolutions

🎨 Advanced Creative Features

  • Template library with pre-built ad formats for different industries
  • Brand kit integration (logos, colors, fonts)
  • Custom character creation and training

🎵 Audio Improvements

  • Voice cloning for brand-specific voiceovers
  • Music library expansion with genre-specific options
  • Sound effect library integration

🤖 AI Enhancements

  • Multi-language support for global audiences
  • Automatic script generation based on product descriptions
  • A/B testing capabilities for different ad variations

📊 Analytics & Optimization

  • Built-in analytics dashboard showing ad performance metrics
  • AI-powered recommendations for improving ad effectiveness
  • Integration with advertising platforms (Google Ads, Facebook Ads)

👥 Collaboration & Platform Expansion

  • Team workspaces for agencies and businesses
  • Mobile app for on-the-go ad creation
  • API access for third-party integrations
  • Marketplace for templates and assets

💰 Monetization & Enterprise

  • Freemium model with premium features
  • Subscription tiers for different use cases
  • Enterprise solutions with custom AI model training and dedicated support

🎯 Vision AdsGenie aims to become the go-to platform for video ad creation, democratizing access to professional-quality advertisements. We envision a future where anyone with a creative idea can bring it to life, regardless of their technical skills or budget. By continuously improving our AI capabilities and expanding our feature set, we're building toward a world where creating stunning video ads is as simple as describing your vision.

AdsGenie: Where Every Creative Wish Becomes Reality ✨

Built With

Share this project:

Updates