🔍 Ollama Web Search

A sophisticated AI-powered web search assistant that combines the intelligence of Ollama with comprehensive web search capabilities.

✨ Features

• 🧠 AI-Powered Query Optimization - Automatically converts natural language questions into effective search queries
• 🌐 Multiple Search Engine Fallbacks - Uses multiple SearxNG instances for reliable search results
• 📄 Smart Content Extraction - Retrieves and processes webpage content using Jina Reader API
• 💾 Search History Tracking - Automatically saves and manages your search history
• 🎨 Beautiful Terminal Interface - Rich colors and formatting for excellent user experience
• ⚡ Streaming Responses - Real-time AI responses with typing effect
• 🔧 Configurable Settings - Easy customization through config.json
• 🛡️ Robust Error Handling - Graceful handling of network issues and API failures

🚀 Quick Start

Installation

1. Clone the repository:

git clone https://github.com/garyku0/ollama-web-search.git
cd ollama-web-search

2. Run the setup script:

python3 setup.py

3. Make sure Ollama is running:

ollama serve

Usage

Interactive Mode (Recommended):

python3 main.py

View Search History:

python3 main.py --history

View Configuration:

python3 main.py --config

Single Query Mode:

python3 main.py --query "What is Python programming?"

Use Different Model:

python3 main.py --model llama2

Get Help:

python3 main.py --help

💡 Example Questions to Try

• "What are the latest developments in artificial intelligence?"
• "How do I install Docker on macOS?"
• "Best practices for Python web development 2025"
• "What is the difference between React and Vue.js?"
• "How to optimize PostgreSQL performance?"

⚙️ Configuration

Edit config.json to customize settings:

{
  "model": "llama3", // Ollama model to use
  "searxng_instances": [
    // Search engines (fallback order)
    "https://search.inetol.net/search",
    "https://searx.be/search"
  ],
  "max_results": 8, // Number of search results to process
  "timeout": 10, // Request timeout in seconds
  "max_retries": 3, // Retry attempts for failed requests
  "history_file": "search_history.json", // Search history storage
  "enable_colors": true, // Terminal colors
  "streaming_delay": 0.02 // Typing effect speed
}

🔧 Troubleshooting

"Cannot connect to Ollama"

# Make sure Ollama is running
ollama serve

# Check if your model is available
ollama list

# Pull the model if needed
ollama pull llama3

"All search instances failed"

• Check your internet connection
• Try updating SearxNG instances in config.json
• Some instances may be temporarily down

Colors not showing properly

• Set "enable_colors": false in config.json
• Some terminals don't support ANSI colors

📊 Search History

Your searches are automatically saved to search_history.json:

• View with python3 main.py --history
• Or type history in interactive mode
• Automatically keeps last 50 searches
• Includes timestamps and results

🔒 Privacy & Security

• Local Processing: All AI processing happens locally via Ollama
• No API Keys: No external AI service API keys required
• Search Privacy: Uses privacy-focused SearxNG instances
• Data Control: All data stays on your machine

🛠️ Requirements

• Python 3.7+
• Ollama (running locally)
• Internet connection
• Required packages (installed via setup.py):
  • ollama>=0.2.0
  • requests>=2.31.0

💡 Tips for Best Results

1. Be Specific: More detailed questions get better results
2. Use Keywords: Include relevant technical terms
3. Ask Follow-ups: Use the interactive mode for related questions
4. Check History: Review past searches to avoid duplicates
5. Experiment with Models: Try different Ollama models for variety

🤝 Contributing

Feel free to open issues or submit pull requests for:

• Bug fixes
• New features
• Documentation improvements
• Additional SearxNG instances
• UI/UX enhancements

📄 License

This project is licensed under the MIT License.