AI Ad Generator automatically creates professional social media advertisements from any website URL. Using advanced multi-agent AI architecture, it analyzes your landing page, develops creative strategies, and generates Instagram-optimized ads with Gemini.

• 🌐 Intelligent Web Analysis: Extracts brand information using BrowserBase Stagehand
• 🎯 Instagram-Optimized Output: Generates square (1:1) creatives tailored for the Instagram feed
• ⚡ Automated Workflow: Coordinates analysis, strategy, and production end-to-end
• 🧠 Strategic AI: Develops data-driven creative briefs and strategies
• 🎨 Gemini Image Generation: High-quality AI-generated visuals with reference screenshots

• Node.js 20+
• Git
• OpenAI API Key - Get yours here
• BrowserBase Account - Sign up here

# Clone the repository
git clone <your-repo-url>
cd ai-ad-generator

# Install dependencies
npm install

# Copy environment variables
cp .env.example .env

Edit .env file with your API keys:

# Required
OPENAI_API_KEY=your_openai_api_key_here
BROWSERBASE_API_KEY=your_browserbase_api_key_here
BROWSERBASE_PROJECT_ID=your_browserbase_project_id_here

# Optional - VoltOps Platform
# Get your keys at https://console.voltagent.dev/tracing-setup
VOLTAGENT_PUBLIC_KEY=your-public-key
VOLTAGENT_SECRET_KEY=your-secret-key

# Server Configuration
PORT=3000

# Development mode (with hot reload)
npm run dev

# The server will start on http://localhost:3000

This application uses VoltAgent's multi-agent architecture:

1. Landing Page Analyzer - Web scraping and brand extraction
2. Ad Creator - Synthesizes strategy and generates Instagram-ready creatives
3. Supervisor - Orchestrates the workflow between agents

• VoltAgent: Multi-agent AI framework
• BrowserBase Stagehand: Intelligent web automation
• OpenAI GPT-4: Agent intelligence
• Google Gemini 2.5 Image: Image generation
• TypeScript: Type-safe development

curl -X POST http://localhost:3000/api/agents/AdGenerationSupervisor/generate-text \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Generate Instagram ads for https://example.com",
    "conversationId": "ad-session-1"
  }'

curl -X POST http://localhost:3000/api/workflows/ad-generation/execute \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "url": "https://example.com",
      "platforms": ["instagram"]
    }
  }'

curl -X POST http://localhost:3000/api/agents/AdGenerationSupervisor/stream-text \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Analyze and generate ads for https://example.com",
    "conversationId": "ad-session-1"
  }'

The VoltOps Platform provides real-time observability for your agents during development:

1. Start your agent: Run npm run dev
2. Open console: Visit console.voltagent.dev
3. Auto-connect: The console connects to your local agent at http://localhost:3141

Features:

• 🔍 Real-time execution visualization
• 🐛 Step-by-step debugging
• 📊 Performance insights
• 💾 No data leaves your machine

For production environments, configure VoltOpsClient:

1. Create a project: Sign up at console.voltagent.dev/tracing-setup
2. Get your keys: Copy your Public and Secret keys
3. Add to .env:

   VOLTAGENT_PUBLIC_KEY=your-public-key
   VOLTAGENT_SECRET_KEY=your-secret-key

4. Configure in code: The template already includes VoltOpsClient setup!

ai-ad-generator/
├── src/
│   ├── index.ts                    # Main VoltAgent configuration
│   ├── lib/
│   │   └── stagehand-manager.ts   # BrowserBase session management
│   ├── agents/                     # Specialized AI agents
│   │   ├── landing-page-analyzer.agent.ts
│   │   ├── ad-creator.agent.ts
│   │   └── supervisor.agent.ts
│   ├── tools/                      # Agent tools
│   │   ├── browserbase/            # Web automation tools
│   │   │   ├── page-navigate.tool.ts
│   │   │   ├── page-extract.tool.ts
│   │   │   ├── page-observe.tool.ts
│   │   │   ├── page-act.tool.ts
│   │   │   └── screenshot.tool.ts
│   │   └── image-generation/       # Ad generation tools
│   │       └── instagram-ad-gemini.tool.ts
│   └── workflows/                  # Workflow definitions
│       └── ad-generation.workflow.ts
├── output/                         # Generated ads directory
│   ├── screenshots/
│   └── ads/
│       └── instagram/
├── .env                           # Environment variables
├── .voltagent/                    # Agent memory storage
└── package.json

output/
├── screenshots/
│   └── screenshot_1234567890.png      # Landing page screenshot
├── ads/
│   └── instagram/
│       └── instagram_brandname_1234567890.png

{
  "brandAnalysis": {
    "productName": "TechStart Pro",
    "tagline": "Innovate Faster, Scale Smarter",
    "valueProposition": "All-in-one platform for startups",
    "targetAudience": "Tech entrepreneurs and small teams"
  },
  "generatedAds": [
    {
      "platform": "instagram",
      "filepath": "output/ads/instagram/instagram_techstart_pro_1234567890.png",
      "style": "primary"
    }
  ],
  "summary": {
    "totalAdsGenerated": 1,
    "platforms": ["instagram"],
    "notes": "Single high-converting Instagram asset"
  }
}

Build and run with Docker:

# Build image
docker build -t ai-ad-generator .

# Run container
docker run -p 3141:3141 --env-file .env ai-ad-generator

# Or use docker-compose
docker-compose up

• npm run dev - Start development server with hot reload
• npm run build - Build for production
• npm start - Run production build
• npm run lint - Check code quality
• npm run lint:fix - Auto-fix linting issues
• npm run typecheck - Run TypeScript type checking

Create a new tool in src/tools/image-generation/:

import { createTool } from "@voltagent/core";
import { z } from "zod";

export const generateLinkedInAdTool = createTool({
  name: "generate_linkedin_ad",
  description: "Generate LinkedIn ad (1200x627)",
  parameters: z.object({
    productName: z.string(),
    tagline: z.string(),
    adConcept: z.string(),
    professional: z.boolean().default(true),
  }),
  execute: async ({ productName, tagline, adConcept }) => {
    // LinkedIn-specific ad generation logic
    // Return generated ad details
  },
});

Modify agent instructions in src/agents/:

export const createAdCreatorAgent = (memory: Memory) => {
  return new Agent({
    name: "AdCreator",
    instructions: `Your custom instructions here...`,
    // Add your custom tools
    tools: [generateLinkedInAdTool, ...otherTools],
    memory,
  });
};

1. BrowserBase Connection Errors

   • Verify your API key and project ID are correct
   • Check BrowserBase service status
   • Ensure you have available browser sessions in your plan

2. OpenAI Rate Limits

   • Implement exponential backoff for retries
   • Consider upgrading your OpenAI plan
   • Reduce the number of parallel requests

3. Memory Database Issues

   # Clear the database and start fresh
   rm -rf .voltagent/*.db

4. Image Generation Failures

   • Check DALL-E API quota
   • Verify output directories have write permissions
   • Ensure prompts comply with OpenAI content policy

• VoltAgent Documentation: voltagent.dev/docs
• BrowserBase Docs: docs.browserbase.com
• OpenAI API: platform.openai.com/docs
• Discord Community: Join VoltAgent Discord
• Example Projects: github.com/VoltAgent/voltagent/tree/main/examples

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

MIT License - see LICENSE file for details

• Built with VoltAgent - TypeScript framework for AI agents
• Powered by BrowserBase - Cloud browser infrastructure
• Image generation by OpenAI DALL-E 3
• Inspired by the Python browser-use ad generator example