# AI Chat Setup Guide

The University Portal includes an AI-powered chat assistant that can help users with information about the university, programs, admissions, and more. The AI chat is powered by Ollama, a local AI model runner.

## Features

- 🤖 **AI-Powered Assistant**: Get instant answers about university information
- 💬 **Real-time Chat**: Interactive conversation interface
- 🏠 **Local Processing**: Runs locally on your machine for privacy
- 📱 **Responsive Design**: Works on desktop and mobile devices
- 🎯 **Context-Aware**: Specialized for university-related questions

## Quick Setup

### 1. Install Ollama

Run the automated setup script:

```bash
npm run setup:ollama
```

This script will:
- Install Ollama on your system
- Start the Ollama service
- Download the Llama2 model

### 2. Manual Installation (Alternative)

If the automated script doesn't work, install Ollama manually:

#### macOS/Linux
```bash
curl -fsSL https://ollama.ai/install.sh | sh
```

#### Windows
Download from: https://ollama.ai/download

### 3. Start Ollama

```bash
ollama serve
```

### 4. Download a Model

```bash
ollama pull llama2
```

### 5. Start the Development Server

```bash
npm run dev
```

## Using the AI Chat

1. **Access the Chat**: Click the AI chat button in the bottom-right corner of any page
2. **Ask Questions**: Type your questions about:
   - University programs and courses
   - Admission requirements
   - Campus life and facilities
   - Research opportunities
   - Student services
3. **Get Instant Answers**: The AI will provide helpful, contextual responses

## Available Models

The default model is `llama2`, but you can use any model supported by Ollama:

```bash
# List available models
ollama list

# Pull a different model
ollama pull codellama
ollama pull mistral
ollama pull llama2:13b
```

## Configuration

### Environment Variables

You can configure the Ollama connection in your `.env.local` file:

```env
OLLAMA_HOST=http://localhost:11434
```

### Model Selection

To use a different model, modify the API call in `src/app/api/chat/ollama/route.ts`:

```typescript
body: JSON.stringify({
  message: userMessage.text,
  model: 'codellama' // Change this to your preferred model
}),
```

## Troubleshooting

### Ollama Not Running

If you see "Ollama is not running" in the chat:

1. Check if Ollama is installed:
   ```bash
   ollama --version
   ```

2. Start the Ollama service:
   ```bash
   ollama serve
   ```

3. Verify it's running:
   ```bash
   curl http://localhost:11434/api/tags
   ```

### Model Not Found

If you get a model error:

1. List available models:
   ```bash
   ollama list
   ```

2. Pull the required model:
   ```bash
   ollama pull llama2
   ```

### Performance Issues

- **Slow Responses**: Try a smaller model like `llama2:7b`
- **High Memory Usage**: Close other applications or use a smaller model
- **Network Issues**: Ensure Ollama is running locally

## Commands Reference

```bash
# Start Ollama service
ollama serve

# Stop Ollama service
pkill ollama

# List models
ollama list

# Pull a model
ollama pull <model-name>

# Run a model interactively
ollama run llama2

# Remove a model
ollama rm <model-name>
```

## Security & Privacy

- **Local Processing**: All AI processing happens locally on your machine
- **No Data Collection**: No user data is sent to external servers
- **Model Control**: You control which models are installed and used

## Support

If you encounter issues:

1. Check the browser console for errors
2. Verify Ollama is running: `curl http://localhost:11434/api/tags`
3. Check the model is installed: `ollama list`
4. Restart Ollama: `pkill ollama && ollama serve`

## Advanced Usage

### Custom System Prompts

You can customize the AI's behavior by modifying the system prompt in `src/app/api/chat/ollama/route.ts`:

```typescript
const systemPrompt = `You are a helpful AI assistant for a university portal...`;
```

### Multiple Models

You can implement model selection by adding a dropdown in the chat interface and passing the selected model to the API.

### Streaming Responses

For real-time responses, you can implement streaming by modifying the API to use Ollama's streaming capabilities.