Walls - AI Assistant Ecosystem

A modular AI assistant ecosystem with voice/text interfaces, RAG capabilities, command execution, and integrated applications for radio streaming, web browsing, and document editing.

A comprehensive AI assistant system with voice and text interfaces, featuring RAG (Retrieval-Augmented Generation), command execution, and integrated applications.

🌟 Overview

Walls is a modular AI assistant ecosystem that combines multiple applications and services to provide a seamless AI-powered experience. The system features both voice and text-based interactions, intelligent command execution, and integrated applications for radio streaming, web browsing, and document editing.

🏗️ Architecture

Core Components

AI Interface (ai_interface/) - Main GUI application with text and voice modes
RAG System (rag/) - Retrieval-Augmented Generation for intelligent responses
Shared Server (shared_server/) - Inter-application communication hub
GUI Core (gui_core/) - Reusable UI components and theming

Integrated Applications

Radio Player (radio_player/) - Internet radio streaming with modern GUI
Browser (browser/) - Web browser with AI integration
Words (Words/) - Document editor with Python GUI and Rust core

🚀 Features

AI Interface

Dual Mode Operation: Text-based chat interface and voice interaction
Command Execution: Automatic detection and execution of bash commands from AI responses
RAG Integration: Context-aware responses using document retrieval
Real-time Streaming: Live AI response streaming with loading indicators
Chat Management: Persistent chat sessions with history

Voice Mode

Wake Word Detection: "Hey Assistant" activation
Speech Recognition: Vosk-powered offline speech-to-text
Text-to-Speech: Kokoro TTS for natural voice responses
Audio Processing: Real-time audio capture and processing

Command Interception

Smart Detection: Automatic bash code block recognition in AI responses
Safety Policies: Blacklist-based command filtering for security
Application Auto-start: Automatic launching of required services
Command Corrections: Intelligent error correction for common mistakes

Radio Player

Internet Radio: Stream thousands of radio stations worldwide
Modern GUI: Beautiful, responsive interface with visualizations
CLI Interface: Command-line control for automation
Search & Discovery: Find stations by genre, country, or name
Metadata Display: Real-time song information and station details

Browser Integration

AI-Powered Browsing: Intelligent web navigation
Content Summarization: AI-powered page summaries
Bookmark Management: Organized bookmark system
Headless Mode: Background web operations

Document Editor (Words)

Hybrid Architecture: Python GUI with Rust core for performance
Rich Text Editing: Advanced document editing capabilities
AI Integration: Smart writing assistance

📦 Installation

Prerequisites

Python 3.8+
Virtual environment support
macOS (primary platform)

Setup

Clone the repository:

git clone https://github.com/askscience/Walls.git
cd Walls

Create virtual environment:

python -m venv venv
source venv/bin/activate  # On macOS/Linux

Install dependencies:
```
pip install -r requirements.txt
```
Install Kokoro TTS (for voice mode):
```
./install_kokoro.sh
```
Start all services:
```
python start_all.py
```

🎯 Usage

Starting the System

All Services:

python start_all.py

Individual Components:

# AI Interface (Text Mode)
python ai_interface/main.py

# Voice Mode
python ai_interface/voice_mode/voice_manager.py

# Radio Player
python -m radio_player.modern_gui

# Browser
python -m browser.main

# Shared Server
python shared_server/start_server.py

AI Interface Usage

Text Mode: Type questions and receive AI responses with automatic command execution
Voice Mode: Say "Hey Assistant" followed by your question
Command Execution: AI responses containing bash code blocks are automatically executed
Radio Control: Ask for music and the system will search and play radio stations

Voice Commands Examples

"Hey Assistant, play some jazz music"
"Hey Assistant, search for news radio stations"
"Hey Assistant, open the browser"
"Hey Assistant, what's the weather like?"

Text Interface Examples

"Find and play classical music stations"
"Search for radio stations in France"
"Open the word editor"
"Browse to wikipedia.com"

🔧 Configuration

Environment Variables

AI_IFACE_DEBUG=1 - Enable debug logging for command interception
RAG_DEBUG=1 - Enable RAG system debug output
VOICE_DEBUG=1 - Enable voice processing debug logs

Configuration Files

rag/config.py - RAG system configuration
rag/prompts.json - AI prompt templates
shared_server/config.py - Inter-service communication settings

🛠️ Development

Project Structure

Walls/
├── ai_interface/          # Main AI interface application
│   ├── components/        # UI components
│   ├── main/             # Main interface logic
│   ├── services/         # Background services
│   ├── utils/            # Utilities and command handling
│   └── voice_mode/       # Voice interaction components
├── rag/                  # RAG system
├── shared_server/        # Inter-application communication
├── gui_core/            # Reusable UI components
├── radio_player/        # Internet radio application
├── browser/             # Web browser integration
├── Words/               # Document editor
└── requirements.txt     # Python dependencies

Key Technologies

GUI Framework: PySide6 (Qt)
AI/ML: Transformers, sentence-transformers, FAISS
Voice Processing: Vosk (STT), Kokoro (TTS)
Web: Requests, BeautifulSoup
Audio: PyAudio, sounddevice
Database: Vector stores for RAG

Adding New Features

New Commands: Add patterns to ai_interface/utils/command_corrections.py
UI Components: Extend gui_core/components/
Voice Commands: Modify voice processing in ai_interface/voice_mode/
RAG Data: Add documents to rag/data/

🔒 Security

Command Execution Safety

Blacklist Filtering: Dangerous commands are automatically blocked
Sandboxed Execution: Commands run in controlled environment
User Confirmation: Critical operations require confirmation
Audit Logging: All command executions are logged

Safe Commands

Application launches (radio_player, browser, etc.)
File operations within project directory
Network requests to known APIs
System information queries

Blocked Commands

System modification commands (rm -rf, sudo, etc.)
Network security bypasses
Privilege escalation attempts
Destructive file operations

🐛 Troubleshooting

Common Issues

Voice Recognition Not Working:

Check microphone permissions
Verify Vosk models are installed
Enable voice debug logging

Commands Not Executing:

Enable AI_IFACE_DEBUG=1
Check command safety policies
Verify virtual environment activation

Radio Player Issues:

Check internet connection
Verify shared server is running
Try restarting the radio player service

Import Errors:

Ensure virtual environment is activated
Check all dependencies are installed
Verify Python path configuration

🙏 Credits and Acknowledgments

Third-Party Components

ollama-mcp-bridge: Originally created by jonigl. This component provides an API layer in front of the Ollama API, seamlessly adding tools from multiple MCP servers. We've included a modified version with local bugfixes in this repository.

📝 License

This project is licensed under the GPL-3 License. See individual component LICENSE files for specific details.

🤝 Contributing

Fork the repository
Create a feature branch
Make your changes
Add tests if applicable
Submit a pull request

📞 Support

For issues and questions:

Check the troubleshooting section
Enable debug logging for detailed error information
Review component-specific README files
Submit issues with detailed logs and reproduction steps

Walls - Your intelligent AI assistant ecosystem, bringing together voice, text, and application integration in one powerful platform.