A web application that allows users to chat with Ollama's LLM models through a user-friendly interface, with all conversations stored in a MySQL database.

• Real-time chat with Ollama models through a web interface
• Intelligent hashtag generation based on conversation content (with enhanced entity extraction)
• MySQL database for persistent storage of conversations (developed on MariaDB)
• Configuration management for model selection and parameters
• Debugging utilities for troubleshooting

• Node.js 14+ and npm
• MySQL database (or MariaDB)
• Ollama installed and running (https://localhost:11434)

1. Clone the repository:

   git clone https://your-repository-url/ollama-chat-web.git
   cd ollama-chat-web
   

2. Install dependencies:

   npm install
   

3. Create a .env file in the project root with the following content:

   DB_HOST=localhost
   DB_USER=your_database_user
   DB_PASSWORD=your_database_password
   DB_NAME=ollama_chat
   PORT=3000
   

   note: 3305 for MariaDb 5.0 3307 for MariaDb 10.0

4. Set up the database:

   npm run db-setup
   

1. Start the application:

   npm start
   

2. Access the web interface at http://localhost:3000 or 3305 or 3307 (see .env creation above)

3. In the settings modal:

   • Select your preferred Ollama model
   • Adjust temperature and token settings
   • Configure context message count

project-root/
├── server.js              // Main entry point
├── db.js                  // Database module
├── db-setup.js            // Database setup script
├── db-test.js             // Database testing utility
├── ollama-direct.js       // Direct Ollama API testing
├── config/
│   └── config-manager.js  // Configuration loading and management
├── routes/
│   ├── api-routes.js      // API endpoints
│   └── debug-routes.js    // Debugging routes
├── services/
│   ├── hashtag-service.js // Enhanced hashtag generation
│   └── ollama-service.js  // Ollama API interaction
├── socket/
│   └── socket-handler.js  // Socket.IO connection handling
└── public/                // Frontend assets and client-side code
    ├── index.html
    ├── client.js
    └── styles.css

The system features an improved hashtag generation algorithm that extracts specific entities from conversations and categorizes them based on the 5W's (Who, What, Where, When, Why/How):

• People/Organizations (Who) - Extracts names and organizations using capitalization patterns
• Actions/Verbs (What) - Identifies important verbs that describe key activities
• Locations (Where) - Recognizes place names and locations
• Time References (When) - Captures dates, months, and temporal expressions
• Concepts/Technical terms (Why/How) - Identifies domain-specific technical vocabulary

This categorization provides more meaningful and specific hashtags for better conversation organization and searchability.

• /api/config - Get/set configuration
• /api/models - List available Ollama models
• /api/check-model/:modelName - Check if a model is installed
• /api/set-model - Directly set the model
• /api/tags - Get all conversation tags
• /api/conversations - Get recent conversations
• /api/search?tag=hashtag - Search conversations by tag

• Visit /debug/database for a visual database browser
• Use /api/run-diagnostics for system diagnostic information
• Check server logs for detailed request/response tracking