Skip to content

README.md

Latest commit

 

History

History
368 lines (265 loc) · 7.38 KB

README.md

File metadata and controls

368 lines (265 loc) · 7.38 KB

🐧 IncognitoAI - Linux/macOS Setup Guide(WORKS WINDOWS RUN bot.py)

A complete Linux/macOS setup for IncognitoAI with both Streamlit and Flask-based Cyberpunk interfaces.

📋 Contents

  • setup.sh - Automated Linux setup script (run this first!)
  • start.sh - Launch Streamlit version
  • bot.sh - Launch Flask Cyberpunk version
  • bot.py - Flask-based cyberpunk AI interface
  • templates/ - HTML templates for Flask app
  • static/ - CSS and JavaScript for Flask app

🚀 Quick Start

Option 1: Automated Setup (Recommended) ⭐

cd IncognitoAI/linux
chmod +x setup.sh
./setup.sh

The setup script will:

  • ✅ Check/install Python 3.8+
  • ✅ Create virtual environment
  • ✅ Install all Python dependencies
  • ✅ Check/install Ollama
  • ✅ Verify Ollama models

Option 2: Manual Setup

# Navigate to linux directory
cd IncognitoAI/linux

# Create virtual environment
python3 -m venv venv
source venv/bin/activate

# Install dependencies
pip install --upgrade pip
pip install -r ../requirements.txt

# Pull AI models
ollama pull llama3.2:1b
ollama pull all-minilm

⚙️ Prerequisites

  • Linux/macOS (Ubuntu 20.04+, Debian, Fedora, macOS 10.14+, etc.)
  • Python 3.8+
  • Ollama (installed automatically or manually)
  • Internet connection (for initial model download only)

📦 System Dependencies

The setup script will auto-install:

  • python3 - Python interpreter
  • python3-pip - Python package manager
  • python3-venv - Virtual environment tool
  • curl - For downloading Ollama

🎯 Running the Apps

1️⃣ Start Ollama (Required in a separate terminal)

ollama serve

You should see: Listening on 127.0.0.1:11434

2️⃣ Start IncognitoAI (Choose one)

Streamlit Version (Clean Web UI):

cd IncognitoAI/linux
chmod +x start.sh
./start.sh
  • Opens at http://localhost:8501
  • Clean, simple interface
  • Perfect for getting started

Flask Cyberpunk Version (Modern Neon UI):

cd IncognitoAI/linux
chmod +x bot.sh
./bot.sh
  • Opens at http://localhost:5000
  • Dark neon cyberpunk aesthetic
  • Modern, sleek design
  • RAG mode toggle

🦙 Ollama Installation

Automatic (via setup.sh)

The setup script handles this for you.

Manual Installation

Linux:

curl -fsSL https://ollama.ai/install.sh | sh

macOS:

# Download from https://ollama.ai/download/mac
# Or use Homebrew
brew install ollama

Then start Ollama:

ollama serve

📊 Using IncognitoAI

Upload and Chat with Documents (RAG Mode)

  1. Start both Ollama and IncognitoAI (see above)
  2. Upload a document:
    • Supported formats: PDF, TXT, Markdown
  3. Enable RAG Mode in the UI
  4. Ask questions about your document

The AI will use your documents for context-aware responses.

Example Questions

  • "Summarize this document"
  • "What are the main points?"
  • "Find information about [topic]"

🎨 Flask Cyberpunk Features

The bot.py Flask app includes:

  • 🌐 Modern web interface with WebSocket support
  • 💜 Dark neon cyberpunk aesthetic
  • ⚡ Real-time streaming responses
  • 📁 Document upload (RAG integration)
  • 🔄 RAG mode toggle
  • 🟢 Live system status indicator
  • 📱 Responsive design for mobile/tablet
  • 🎯 Command-line style chat UI

🔧 Troubleshooting

Error: "Ollama not found" or "Connection refused"

Solution 1: Install Ollama manually

curl -fsSL https://ollama.ai/install.sh | sh

Solution 2: Make sure Ollama is running in another terminal

ollama serve

Error: "Python not found"

# Ubuntu/Debian
sudo apt-get update
sudo apt-get install python3 python3-pip

# macOS
brew install python@3.11

Error: "Permission denied" on scripts

chmod +x setup.sh start.sh bot.sh

Error: "Port 8501 already in use" (Streamlit)

# Run on a different port
streamlit run ../app.py --server.port 8502

Error: "Port 5000 already in use" (Flask)

Edit bot.py and change the port (search for app.run):

app.run(debug=True, host='0.0.0.0', port=5001)

Error: "Models failed to load"

Pull models manually:

ollama pull llama3.2:1b
ollama pull all-minilm

Error: "Virtual environment not found"

Run setup again:

./setup.sh

⚙️ Configuration

Change AI Models

Edit bot.py (lines 10-12):

MODEL_NAME = "llama2:7b"              # Change to your model
EMBEDDING_MODEL = "nomic-embed-text"  # Change to your embedding model

View available models:

ollama list
ollama pull [model-name]

Change Flask Port

Edit bot.py (bottom of file, look for app.run):

app.run(debug=True, host='0.0.0.0', port=5000)  # Change 5000 to desired port

Change Streamlit Settings

Edit or create ~/.streamlit/config.toml:

[server]
port = 8502

🔐 Privacy & Security

  • ✅ 100% Offline - No data leaves your computer
  • ✅ Zero Telemetry - No tracking or analytics
  • ✅ Open Source - Transparent and auditable code
  • ✅ Local Storage - All data stored locally in .chroma_db/
  • ✅ No Registration - Works without any account

📚 Project Dependencies

Key Python packages:

  • flask - Web framework
  • streamlit - UI framework
  • langchain - AI orchestration
  • langchain-ollama - Ollama integration
  • chromadb - Vector database for RAG
  • requests - HTTP client
  • pypdf - PDF parsing

See ../requirements.txt for complete list.

💡 Tips & Best Practices

  1. Two Terminals:

    • Terminal 1: ollama serve (keep running)
    • Terminal 2: ./start.sh or ./bot.sh

  2. Run Both UIs:

    • Terminal 1: ollama serve
    • Terminal 2: ./start.sh (port 8501)
    • Terminal 3: ./bot.sh (port 5000)
    • They share the same AI models and vector database

  3. Use RAG Mode for Best Results:

    • Upload documents first
    • Enable RAG mode
    • Ask specific questions about your documents

  4. Performance Tips:

    • Use llama3.2:1b (fast and efficient)
    • For more power, try llama2:7b (requires more RAM)
    • All-MiniLM embedding model is fast

  5. Document Types Supported:

    • .pdf - PDF files
    • .txt - Plain text
    • .md - Markdown files

🚀 Next Steps

  1. Customize the UI:

    • Modify static/bot.css for Cyberpunk colors
    • Edit templates/bot.html for layout changes

  2. Add Custom Features:

    • Edit bot.py to add endpoints
    • Modify app.py for Streamlit features

  3. Try Different Models:

    ollama pull mistral:latest
ollama pull neural-chat:latest

  4. Improve RAG:

    • Upload large document sets
    • Ask questions about complex topics
    • Use specific terminology

📞 Support & Resources

If Something Breaks

  1. Check logs in terminal output
  2. Verify Ollama is running: curl http://localhost:11434
  3. Rerun setup: ./setup.sh
  4. Check GitHub issues

External Resources

Main Repository

📝 License

MIT License - See ../LICENSE.md for details

🎉 Enjoy your private, offline AI assistant! 🚀