AI-Enhanced Bluesky Feed

Web application with AI-powered image analysis, real-time analytics dashboard, and intelligent reply generation for Bluesky posts. (AKA "Bluesky Reply Guy")

🧠 Concept

• Built to explore GPT-4o's multimodal capabilities (image + text in single API call)
• Focuses exclusively on posts with images/media, ignoring text-only posts
• Combines computer vision with social media interaction
• Real-time image analysis and engagement tracking

🛠️ Tech Stack

• Backend
  • Python Flask
  • AT Protocol/Bluesky API
  • OpenAI GPT-4o API
  • AWS SSM
• Frontend
  • HTML/CSS/JavaScript
  • Bootstrap
  • Server-Sent Events
  • Real-time analytics
• Development
  • Cursor AI agentic coding
  • ChatGPT reasoning assistant
• Features
  • Multimodal analysis
  • Reply analytics
  • Smart response generation

GPT-4o analyzed the image text and incorporated it into the AI-generated response

🚀 Setup

git clone <repository-url>
cd vit-sandbox
pip install -r requirements.txt
python main.py

📡 Key Endpoints

• GET /api/posts - Posts with AI analysis and engagement metrics
• POST /api/ai-reply - Generate contextual replies using multimodal analysis
• GET /api/posts/stream - Real-time stream of AI-analyzed posts

🎯 Project Status

• Sunset: Decided to discontinue after testing
• Reason: Bluesky community values authentic human interaction over AI-generated content
• Learning: Technical capability doesn't always align with user needs

🏗️ Skills Practiced

• Multimodal AI integration with GPT-4o
• AT Protocol/Bluesky API integration
• AI-assisted development workflow

📄 License

MIT License