Markly — Social Bookmarking with AI Suggestions

A social bookmarking built with ASP.NET Core MVC, Entity Framework Core, PostgreSQL with Anthropic tag and category suggestions.

Features

• Accounts, roles, profiles, and social voting/commenting
• Bookmarks with categories, tags, advanced search, and public/private visibility
• AI-assisted tag/category suggestions via Anthropic Claude 4.5 Haiku
• Seed data with ready-to-use accounts for quick demos
• Docker Compose with PostgreSQL and Caddy reverse proxy

Prerequisites

• .NET 9.0 SDK
• Docker & Docker Compose
• PostgreSQL database
• Anthropic API key (for AI suggestions)

Configuration

Database

Configure your PostgreSQL connection string in appsettings.json:

{
  "ConnectionStrings": {
    "DefaultConnection": "Host=localhost;Port=5434;Database=markly;Username=markly_user;Password=markly_password"
  }
}

Caddy (Basic Auth)

Caddy is configured to read the bcrypt hash from the MARKLY_BASICAUTH_HASH environment variable (the project Caddyfile uses {$MARKLY_BASICAUTH_HASH}).

Generate a bcrypt hash for the password (example uses ImiPlaceASPdotnet):

python3 -c 'import bcrypt; print(bcrypt.hashpw(b"ImiPlaceASPdotnet", bcrypt.gensalt()).decode())'

Set the environment variable before starting Caddy:

export MARKLY_BASICAUTH_HASH="$2b$12$..."

AI Suggestions (Anthropic Claude)

The application uses Claude 4.5 Haiku for AI-powered tag and category suggestions. To enable this feature:

1. Get an API key from Anthropic Console

2. Configure the API key using one of these methods:

   Option A: User Secrets (Recommended for development)

   dotnet user-secrets init
   dotnet user-secrets set "Anthropic:ApiKey" "your-api-key-here"

   Option B: Environment Variable

   export Anthropic__ApiKey="your-api-key-here"

   Option C: appsettings.json (Not recommended for production)

   {
     "Anthropic": {
       "ApiKey": "your-api-key-here",
       "Model": "claude-haiku-4-5",
       "MaxTokens": 256
     }
   }

3. Rate Limiting Configuration (optional):

   {
     "RateLimiting": {
       "MaxRequestsPerWindow": 10,
       "WindowSeconds": 60
     }
   }

   This limits each user to 10 AI suggestion requests per minute.

Running the Application

Option 1: Local Development

Tip: You can use Docker to run just the PostgreSQL database while developing locally:

docker compose up -d postgres

This avoids needing to install PostgreSQL locally. The connection string in appsettings.Development.json is already configured to work with this setup.

1. Apply database migrations:

   dotnet ef database update

2. Run the application:

   dotnet run

3. Access the application at https://localhost:5001 or http://localhost:5000

Default Test Accounts (Seed Data)

The database is automatically seeded with sample data and test accounts:

Email	Password	Role	Name
admin@markly.com	Admin123!	Admin	Alex Morgan
sarah@example.com	Sarah123!	User	Sarah Chen
mike@example.com	Mike1234!	User	Michael Johnson

These accounts come with sample bookmarks, categories, tags, comments, and votes. The seed data only runs on an empty database.

Option 2: Docker

1. Configure environment variables in docker-compose.yml:

   environment:
      Anthropic__ApiKey: ${ANTHROPIC_API_KEY}

   You can provide the ANTHROPIC_API_KEY in two common ways:

   • Option A: Export a system environment variable (useful for CI or local shells)

     export ANTHROPIC_API_KEY="your_actual_api_key"
     docker compose up -d

   • Option B: Create a .env file next to docker-compose.yml (recommended for local development)

     ANTHROPIC_API_KEY=your_actual_api_key

2. Start all services (PostgreSQL, app, and Caddy reverse proxy):

   docker compose up -d

3. Access the application at http://localhost (port 80) or https://localhost (port 443)

4. View logs:

   docker compose logs -f markly

5. Stop all services:

   docker compose down

The Docker setup includes:

• PostgreSQL database (data persisted in a Docker volume)
• Markly app with automatic database migrations on startup
• Caddy reverse proxy with automatic HTTPS

Using AI Suggestions

1. Navigate to Create or Edit Bookmark page
2. Enter a title and optionally a description
3. Click the "Suggest with AI" button
4. Review the suggested tags and categories
5. Click on any suggestion to add it to your bookmark
6. New tags/categories will be created automatically if they don't exist

Project Structure

markly/
├── Configuration/          # Settings classes (Anthropic, RateLimiting)
├── Controllers/            # MVC Controllers
├── Data/
│   └── Entities/           # Entity models
├── Helpers/                # Utility classes
├── Migrations/             # EF Core migrations
├── Models/                 # Domain models
├── Services/
│   ├── Interfaces/         # Service contracts
│   └── Implementations/    # Service implementations
├── ViewModels/             # View models
├── Views/                  # Razor views
└── wwwroot/                # Static files (CSS, JS)

Technologies

• ASP.NET Core MVC 9.0
• Entity Framework Core
• PostgreSQL
• ASP.NET Identity
• Bootstrap 5
• Anthropic Claude API

License

This project is for educational purposes.