Skip to content
BranchesTags

Repository files navigation

🤖 Jarvis AI Assistant

A sophisticated AI assistant mobile app inspired by Iron Man's JARVIS, featuring voice recognition, text-to-speech, and integration with local and cloud AI providers including Ollama, OpenAI, DeepSeek, Google AI, OpenRouter, and NVIDIA NIM.

Jarvis AI Assistant

✨ Features

🎙️ Voice Interaction

  • Speech Recognition: Tap to speak and interact naturally
  • Text-to-Speech: Jarvis responds with a sophisticated British accent
  • Auto-send: Voice input automatically processes your requests
  • Voice Visualizer: Real-time audio feedback during conversations

🤖 AI Integration

  • Multiple AI Providers: Choose between Local (Basic), Local (Advance), OpenAI GPT, DeepSeek, Google AI, OpenRouter, or NVIDIA NIM
  • Ollama Integration: Connect to a local Ollama instance through Local (Advance)
  • Automatic Ollama Detection: Detect local Ollama models from the settings panel
  • Test Ollama Connection: Validate the local Ollama endpoint before saving settings
  • Provider Model Detection: Detect available models for OpenAI, DeepSeek, Google AI, OpenRouter, and NVIDIA NIM when a valid API key is provided
  • Custom API Endpoints: Configure provider-specific API URL endpoints directly in settings
  • OpenAI GPT Support: OpenAI-compatible chat completion models
  • DeepSeek Integration: DeepSeek chat models
  • Google AI Integration: Gemini models supported by the configured endpoint
  • OpenRouter Integration: Access to 100+ AI models through one API
  • NVIDIA NIM Integration: Access NVIDIA-hosted chat models through an OpenAI-compatible API
  • Fallback System: Automatically falls back to local responses if API fails
  • Reduced Credential Persistence: API keys are stripped before settings are written back to local storage

📱 Mobile-First Design

  • Progressive Web App (PWA): Install as a native app on any device
  • Local Fallback Mode: Local (Basic) works without a cloud provider
  • iOS Safe Area: Full iPhone/iPad compatibility including notch support
  • Haptic Feedback: Tactile responses for better user experience
  • Touch Optimized: Large buttons and gesture-friendly interface

🎨 Premium UI/UX

  • Jarvis-themed Design: Cyan blue accent colors with dark theme
  • Smooth Animations: Fluid transitions and micro-interactions
  • Responsive Layout: Adapts to all screen sizes
  • Status Indicators: Real-time connection and activity status
  • Advanced Settings Panel: Comprehensive AI and audio configuration

🧠 AI Capabilities

  • Conversational AI: Natural language processing and responses
  • Context Awareness: Intelligent replies based on conversation history
  • Time & Date: Current time and date information
  • Connection Awareness: Local responses can react to online/offline status
  • Professional Personality: British assistant personality like the original JARVIS

🚀 Quick Start

Prerequisites

  • Node.js 18+
  • npm, pnpm, or yarn
  • Ollama (optional, for Local Advance mode)
  • API keys for OpenAI, DeepSeek, Google AI, OpenRouter, or NVIDIA NIM (optional)

Installation

  1. Clone the repository

    git clone https://github.com/inertz/jarvis.git
cd jarvis

  2. Install dependencies

    npm install

  3. Start development server

    npm run dev

  4. Open in browser

    http://localhost:5173

🔑 API Configuration

Local (Advance) / Ollama Setup:

  1. Install and run Ollama
  2. Pull at least one model, for example: 
    ollama pull llama3.2
  3. Open the app settings (gear icon)
  4. Select "Local (Advance)" as your AI provider
  5. Keep the endpoint as http://localhost:11434
  6. Click "Test Ollama Connection"
  7. Choose one of the detected models and save

Notes:

  • Only localhost Ollama endpoints are allowed by the app
  • If Ollama is running, the settings panel will try to detect models automatically
  • Local (Basic) is still available as an offline fallback without Ollama

OpenAI Setup:

  1. Get your API key from OpenAI Platform
  2. Open the app settings (gear icon)
  3. Select "OpenAI GPT" as your AI provider
  4. Enter your API key
  5. Optionally change the API URL Endpoint
  6. Click "Detect Models" to load available models automatically
  7. Choose your preferred model and save

DeepSeek Setup:

  1. Get your API key from DeepSeek Platform
  2. Open the app settings (gear icon)
  3. Select "DeepSeek" as your AI provider
  4. Enter your API key
  5. Optionally change the API URL Endpoint
  6. Click "Detect Models" to load available models automatically
  7. Choose your preferred model and save

Google AI Setup:

  1. Get your API key from Google AI Studio
  2. Open the app settings (gear icon)
  3. Select "Google AI (Gemini)" as your AI provider
  4. Enter your API key
  5. Optionally change the API URL Endpoint
  6. Click "Detect Models" to load available Gemini models automatically
  7. Choose your preferred model and save

OpenRouter Setup:

  1. Get your API key from OpenRouter
  2. Open the app settings (gear icon)
  3. Select "OpenRouter" as your AI provider
  4. Enter your API key
  5. Optionally change the API URL Endpoint
  6. Click "Detect Models" to load available models automatically
  7. Choose your preferred model and save

NVIDIA Setup:

  1. Get your API key from NVIDIA Build / API Catalog
  2. Open the app settings (gear icon)
  3. Select "NVIDIA NIM" as your AI provider
  4. Enter your API key
  5. Optionally change the API URL Endpoint
  6. Click "Detect Models" to load available NVIDIA-compatible models automatically
  7. Choose your preferred model and save

Notes for cloud providers:

  • Model detection depends on the provider allowing browser-side access to its model-list endpoint
  • If a provider blocks the request with CORS or policy restrictions, the app will show the returned error in the settings panel
  • API keys are entered directly in the browser UI, so this setup is best suited to local or personal use unless you add a backend proxy

📲 Install as Mobile App

iPhone/iPad:

  1. Open the app in Safari
  2. Tap the Share button (square with arrow)
  3. Select "Add to Home Screen"
  4. Tap "Add"

Android:

  1. Open the app in Chrome
  2. Tap the menu (3 dots)
  3. Select "Add to Home Screen"
  4. Tap "Add"

🛠️ Built With

  • React 18 - Modern React with hooks and TypeScript
  • TypeScript - Type-safe development
  • Tailwind CSS - Utility-first styling
  • Vite - Fast build tool and dev server
  • PWA - Progressive Web App capabilities
  • Web Speech API - Voice recognition and synthesis
  • Lucide React - Beautiful icons
  • OpenAI API - GPT models integration
  • DeepSeek API - Advanced AI model integration
  • Google AI API - Gemini models integration
  • OpenRouter API - Access to 100+ AI models
  • NVIDIA API - NVIDIA NIM and hosted model access

📁 Project Structure

jarvis-ai-assistant/
├── public/
│   ├── manifest.json          # PWA manifest
│   ├── pwa-*.png             # App icons
│   └── apple-touch-icon.png  # iOS icon
├── src/
│   ├── components/
│   │   └── SettingsPanel.tsx # AI configuration panel
│   ├── services/
│   │   └── aiService.ts      # AI provider integration
│   ├── hooks/
│   │   └── useLocalStorage.ts # Local storage hook
│   ├── types.ts              # TypeScript definitions
│   ├── App.tsx               # Main application component
│   ├── main.tsx              # Application entry point
│   └── index.css             # Global styles
├── index.html                # HTML template
├── vite.config.ts            # Vite configuration
├── tailwind.config.js        # Tailwind CSS config
└── package.json              # Dependencies and scripts

🎯 Usage Examples

Voice Commands

  • "Hello Jarvis" - Greeting and introduction
  • "What time is it?" - Current time
  • "What's the date?" - Current date
  • "How are you?" - System status
  • "Help" - Available commands
  • "Thank you" - Polite acknowledgment

AI Provider Features

  • Local (Basic): Basic built-in responses without a cloud provider
  • Local (Advance): Uses a local Ollama model with automatic model detection
  • OpenAI GPT: Advanced conversational AI with context awareness and model autodetection
  • DeepSeek: Specialized AI with coding and reasoning capabilities and model autodetection
  • Google AI: Powerful Gemini models with advanced reasoning and model autodetection
  • OpenRouter: Access to Claude, GPT, Llama, Gemini, and 100+ other models with model autodetection
  • NVIDIA NIM: Hosted NVIDIA-compatible chat models with model autodetection

Text Input

Type any message in the input field and press Enter or tap the send button.

⚙️ Configuration

AI Settings

  • Provider Selection: Choose between Local (Basic), Local (Advance), OpenAI, DeepSeek, Google AI, OpenRouter, or NVIDIA NIM
  • Ollama Detection: Automatically discover local Ollama models
  • Connection Testing: Test the Ollama endpoint directly from settings
  • Cloud Model Detection: Detect supported models for OpenAI, DeepSeek, Google AI, OpenRouter, and NVIDIA NIM
  • API URL Endpoints: Override the default provider endpoints when needed
  • API Key Management: Show/hide input fields for cloud providers
  • Model Selection: Choose specific models for each provider
  • Fallback System: Automatic fallback to local responses

Audio Settings

  • Voice Output Toggle: Enable/disable text-to-speech
  • British Accent: Automatic selection of British voice
  • Speech Tuning: Speech rate, pitch, and volume are fixed in code

PWA Settings

  • Installable App Shell: Includes a manifest and PWA build configuration
  • Custom App Icons: Professional Jarvis-themed icons
  • Standalone Display: Full-screen app experience
  • Auto-update: Automatic updates when available

🔧 Development

Available Scripts

# Start development server
npm run dev

# Build for production
npm run build

# Preview production build
npm run preview

# Run linting
npm run lint

Environment Setup

No environment variables are required for basic functionality.

Important:

  • Cloud API keys are entered directly in the browser UI
  • The app strips API keys before saving settings back to local storage
  • Because this is a frontend-only app, browser-entered cloud API keys are still exposed to the client runtime
  • Some cloud provider model-detection requests may fail because of CORS or provider-side browser restrictions
  • For stronger security, route provider calls through your own backend

Browser Compatibility

  • Chrome/Edge: Full support including all AI features
  • Safari: Full support (iOS 14.3+) with PWA capabilities
  • Firefox: Limited speech synthesis, full AI support
  • Mobile browsers: Optimized experience with touch controls

💰 API Costs

Pricing changes frequently and depends on the selected provider and model.

Check the provider pricing pages directly before relying on a cost estimate:

🚀 Deployment

Netlify (Recommended)

  1. Build the project: npm run build
  2. Deploy the dist folder to Netlify
  3. Enable HTTPS for PWA features

Vercel

  1. Connect your GitHub repository
  2. Vercel will auto-detect Vite configuration
  3. Deploy with zero configuration

GitHub Pages

  1. Install gh-pages: npm install --save-dev gh-pages
  2. Add deploy script to package.json
  3. Run: npm run deploy

🔒 Security

  • Reduced API Key Persistence: Cloud API keys are stripped before settings are written to local storage
  • No Server Storage: All data remains on your device
  • Local Ollama Restriction: Local (Advance) only allows localhost Ollama endpoints
  • HTTPS Required: PWA features require secure connection
  • Privacy First: No data is sent to third parties except chosen AI providers

🤝 Contributing

  1. Fork the repository
  2. Create a feature branch: git checkout -b feature/amazing-feature
  3. Commit changes: git commit -m 'Add amazing feature'
  4. Push to branch: git push origin feature/amazing-feature
  5. Open a Pull Request

Development Guidelines

  • Follow TypeScript best practices
  • Use Tailwind CSS for styling
  • Maintain mobile-first responsive design
  • Test on multiple devices and browsers
  • Keep accessibility in mind
  • Test with different AI providers

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

  • Inspired by Marvel's Iron Man JARVIS AI
  • Icons by Lucide
  • Built with modern web technologies
  • OpenAI for GPT model access
  • DeepSeek for advanced AI capabilities
  • NVIDIA for hosted model access
  • Special thanks to the React and Vite communities

📞 Support

🔮 Roadmap

  • OpenAI GPT integration
  • DeepSeek AI integration
  • Google AI (Gemini) integration
  • NVIDIA NIM integration
  • Advanced settings panel
  • Local Ollama integration
  • Ollama auto-detection and connection testing
  • Reduced API key persistence
  • OpenRouter integration (100+ models)
  • Backend proxy for secure cloud API key handling
  • Conversation history export
  • Custom voice training
  • Weather and news integration
  • Calendar and reminder features
  • Multi-language support
  • Smart home integration
  • Advanced conversation memory
  • Claude AI integration

About

A sophisticated AI assistant mobile app inspired by Iron Man's JARVIS, featuring voice recognition, text-to-speech, and integration with advanced AI providers like OpenAI and DeepSeek.

Resources

Readme
Activity

Stars

3 stars

Watchers

0 watching

Forks

2 forks
Report repository

Releases

Packages

Contributors

Languages