Jarvis AI Assistant
Your Desktop AI Command Center
Electron desktop AI assistant with wake-word activation (Picovoice Porcupine), ElevenLabs voice synthesis, and access to Claude, Gemini, and 100+ HuggingFace models. Manage your calendar, terminal, files, Telegram, and custom workflows -- all through voice or text. Featuring stunning Three.js 3D visualization.
Wake-Word Activated
Say "Hey Jarvis" -- Picovoice Porcupine listens on-device
Multi-Model AI
Claude, Gemini, HuggingFace, Whisper, and ElevenLabs
Desktop Native
Electron 28 app with 3D visualization and system access
Screenshots
Voice Pipeline
1. Wake Word
Picovoice Porcupine detects "Hey Jarvis" on-device with zero cloud latency
2. AI Processing
Whisper transcribes speech, then Claude/Gemini/HuggingFace processes your request
3. Voice Response
ElevenLabs synthesizes natural voice with streaming audio and 3D visualization
Powerful Features
Wake-Word Detection (Picovoice Porcupine)
Always-listening wake-word detection using Picovoice Porcupine. Say 'Hey Jarvis' and your assistant activates instantly.
- Picovoice Porcupine wake-word engine
- Custom wake-word ('Hey Jarvis')
- Low-latency activation (<200ms)
- Works offline (on-device processing)
- Adjustable sensitivity levels
- Continuous background listening
ElevenLabs Voice Synthesis
Natural, expressive voice responses powered by ElevenLabs TTS with multiple voice profiles and streaming audio.
- ElevenLabs text-to-speech integration
- Multiple voice profiles
- Streaming audio for low latency
- Adjustable speech rate and pitch
- Emotional expression in voice
- Whisper STT for speech recognition
Multi-Model AI Engine
Access Claude (Sonnet/Haiku/Opus), Gemini (2.0 Flash/Pro), HuggingFace (100+ models), and more from one interface.
- Claude Sonnet, Haiku, and Opus
- Gemini 2.0 Flash and Pro
- HuggingFace (100+ open models)
- OpenAI Whisper for transcription
- ElevenLabs for voice synthesis
- Hot-swap between models mid-conversation
Telegram Bot Integration
Access Jarvis from anywhere via Telegram. Send text or voice messages and get AI responses on mobile.
- Telegram Bot API integration
- Text and voice message support
- File and document sharing
- Group chat AI assistance
- Command-based interactions
- Cross-device conversation sync
Google Calendar Integration
Manage your schedule with voice commands. Create events, check availability, and get daily briefings.
- Google Calendar OAuth integration
- Voice-based event creation
- Daily schedule briefings
- Availability checking
- Meeting reminders and alerts
- Multi-calendar support
Shell Terminal Access
Execute terminal commands through voice or text. Run scripts, manage files, and automate system tasks.
- Voice-activated shell commands
- Command output streaming
- Script execution and automation
- Safety guardrails for dangerous commands
- Command history and favorites
- Cross-platform shell support
File Browser & Manager
Browse, search, and manage files through natural language. Open, move, copy, and organize with voice commands.
- Natural language file operations
- File search and discovery
- Drag-and-drop file management
- File preview and quick look
- Bulk operations support
- Recent files and favorites
Custom Workflows
Create automated workflow chains that combine multiple tools, APIs, and AI models into repeatable sequences.
- Visual workflow builder
- Chain multiple tools and APIs
- Conditional logic and branching
- Scheduled workflow execution
- Workflow templates library
- Share workflows across devices
3D Visualization (Three.js)
Interactive 3D AI assistant visualization with Three.js. Dynamic animations respond to voice and conversation state.
- Three.js 3D rendering engine
- Dynamic animation states
- Voice-reactive visualizations
- Conversation state indicators
- Customizable 3D avatar
- GPU-accelerated rendering
100+
HuggingFace Models
5
AI Providers
9
Integrated Tools
3D
Visualization
Tech Stack
AI Models
Meet Your AI Desktop Assistant
Say "Hey Jarvis" and take command of your desktop. Voice-activated AI with Claude, Gemini, 100+ HuggingFace models, and stunning 3D visualization.