All Projects

Jarvis AI Assistant – Cross-Device AI Orchestrator

Cross-device AI assistant orchestrator connecting phone, laptop, desktop, and browser through a FastAPI backend, agent protocol, and LLM-powered orchestration layer. Includes a React/Next.js web and desktop frontend.

AI Systems Engineer2026
GitHub

Stack

PythonFastAPITypeScriptReactNext.jsOpenAI APIWebSocketDocker

Jarvis AI Assistant – Cross-Device AI Orchestrator

Overview

Jarvis is a cross-device AI assistant orchestrator designed to connect multiple devices — phone, laptop, desktop, browser, and future smart devices — through a unified FastAPI backend with an LLM-powered orchestration layer. Unlike tools that require MCP to interact with devices, Jarvis operates independently through its own agent protocol.

Architecture

Jarvis Backend (FastAPI / Python)
├── Agent Protocol      → Device registration, task routing
├── LLM Orchestration   → OpenAI API for intent parsing & reasoning
├── Device Manager      → Cross-device state sync
└── WebSocket Layer     → Real-time device communication
         │
         ▼
Frontend (React / Next.js / TypeScript)
├── Web App             → Browser-based control interface
├── Desktop App         → Electron-based desktop client
└── Mobile Support      → iOS and Android device integration

What I Built

Backend (Python / FastAPI)

  • LLM-powered orchestration layer using OpenAI API for natural language task parsing and multi-step reasoning
  • Agent protocol enabling devices to register, receive tasks, and report results
  • WebSocket-based real-time communication between backend and connected devices
  • Task routing logic that determines which device should handle each request

Frontend (TypeScript / React / Next.js)

  • Web and desktop interface for managing connected devices and issuing commands
  • Real-time device status dashboard showing connected devices and active tasks
  • Natural language input for interacting with the AI orchestration layer

Key Highlights

  • No MCP dependency — custom agent protocol for device interaction
  • Multi-device orchestration through a single AI backend
  • Real-time WebSocket communication for low-latency device control
  • Extensible architecture designed for future smart device integration

Repositories