Jarvis Voice Agent (Under Development)

A real-time AI voice assistant split into two Python services with isolated environments:

• backend: FastAPI REST API for auth, sessions, and LiveKit room management
• worker: LiveKit Agent that runs STT → LLM → TTS for live voice conversations

Project split

• Backend (FastAPI)
  • Auth (Google OAuth + JWT), session/tier control, LiveKit token/room handling
• Worker (LiveKit Agent)
  • Deepgram STT, OpenRouter LLM (e.g., Gemini/GPT), Google Cloud TTS, Silero VAD

Basic structure

jarvis-voice-agent/
├─ backend/
│  ├─ main.py
│  └─ app/{auth,db,routes,schemas,services}
├─ worker/
│  ├─ main.py
│  ├─ entry.py
│  └─ agent.py
└─ google-creds.json (gitignored)

Prerequisites

• Python 3.13+
• uv (Python package manager) installed
• Required service credentials in .env and google-creds.json

Setup and run with uv (isolated per service)

Run each service in its own terminal (separate virtual envs and processes).

• Backend (Terminal 1)

cd backend
uv venv
.venv\Scripts\Activate.ps1
uv sync
uv run uvicorn main:app --reload --port 8000

• Worker (Terminal 2)

cd worker
uv venv
.venv\Scripts\Activate.ps1
uv sync
uv run python main.py dev

Environment

Place environment variables in a .env at the repo root (used by both services):

• DATABASE_URL, UPSTASH_REDIS_REST_URL, UPSTASH_REDIS_REST_TOKEN
• LIVEKIT_API_KEY, LIVEKIT_API_SECRET, LIVEKIT_URL
• JWT_SECRET_KEY, GOOGLE_CLIENT_ID, GOOGLE_CLIENT_SECRET
• DEEPGRAM_API_KEY, OPENROUTER_API_KEY, OPENROUTER_BASE_URL, ELEVENLABS_API_KEY
• GOOGLE_APPLICATION_CREDENTIALS=./google-creds.json

API docs: http://localhost:8000/docs (after backend starts)

Status: actively under development.