📚 Purpose

An AI-powered bookmark management system that lets you save, organize, and search bookmarks using natural language. Built as a showcase for modern serverless AI architecture on AWS.

✨ Features

• Natural Language Interface - Chat with an AI to manage bookmarks
• Canvas Documents - Rich document tabs for agent-generated guides, tutorials, and reports with edit, copy, and download
• Smart Categorization - Auto-categorize URLs using AI analysis
• URL Analysis - Extract metadata and summaries automatically
• Reading Lists - Organize bookmarks into reading lists with read/unread tracking
• 18 Categories - AWS, AI/ML, Python, Tools, and more
• File Upload - Attach JSON, text, PDF, and images for agent processing
• Real-time Streaming - HTTP SSE streaming responses from the agent
• Public/Private - Control visibility of your bookmarks

🏗 Architecture

┌─────────────────────────────────────────────────────────┐
│                      Frontend                            │
│   CloudFront → S3 + API Gateway + WAF                   │
│   Lambda Function URL → SSE Streaming                   │
└──────────────────────────┬──────────────────────────────┘
                           │
                           ▼
┌─────────────────────────────────────────────────────────┐
│                  AgentCore Runtime                       │
│   Strands Agent + Claude 4.5 + Bedrock Guardrails       │
│                         │                                │
│   ┌─────────────────────┼─────────────────────┐         │
│   │                     │                     │         │
│   ▼                     ▼                     ▼         │
│ Runtime Tools     AgentCore Gateway     AgentCore Memory│
│ (9 tools)         (10 DynamoDB tools)   (conversations) │
└─────────────────────────────────────────────────────────┘
                        

🛠 Tech Stack

💰 Estimated Monthly Cost

Based on moderate usage (~1,000 requests/month)

* LLM inference is ~60% of cost. Lambda, CloudFront, Cognito covered by free tier.

🔧 Agent Tools (19 total)

💻 Source Code

This project is open source and available on GitHub: