Jetson-first companion runtime

NodeSpark Synra

A standalone AI companion built from Synra's NodeSpark identity: voice, avatar presence, local or cloud models, memory, free smart-home tools, and premium NodeSpark Command Center skills.

Install on Jetson Contribute Runbook Roadmap

Synra Classic waving in the Orbit Lounge companion runtime — **Synra Classic** Live VRM companion with control panels, local model status, and NodeSparkHub access. 4.3

Built to feel alive

Synra combines a full-screen companion stage, visible runtime controls, appliance-grade settings, and timestamped voice motion for a smarter Jetson assistant experience.

Synra Classic standing in the Orbit Lounge live companion view — **Live companion stage** Orbit Lounge background, Synra Classic, VRMA motion controls, and local model connection status.

Synra settings panel with Home Assistant configuration — **Appliance control center** AI, voice, memory, Home Assistant, NodeSparkHub, display, backups, and restore settings in one place.

Synra Classic waving with the live control panels visible — **Motion and presence** Synra Classic remains the main avatar, with 97 VRMA motion clips and Jetson-friendly rendering.

Alive, But Reliable

Synra keeps clear visual states for idle, listening, thinking, speaking, and offline moments so users know what just happened.

Local First

The Jetson path runs a lean local server, supports OpenAI-compatible model endpoints, and fails safely when a model or tool is not configured.

Permissioned Tools

Camera, memory, smart-home, and future NodeSpark skills are explicit and user controlled, with confirmation gates for important actions.

Production kiosk path

Install Synra on Jetson

The installer sets up the Synra web runtime, local API service, private model configuration, and the dedicated Electron kiosk shell that avoids snap Chromium confinement.

curl -fsSL https://raw.githubusercontent.com/synryzen/NodeSpark-Synra/main/scripts/install-jetson.sh | bash

1. Install Builds Synra, installs the user service, and creates a safe starter config.

2. Configure AI Edit ~/.config/synra-standalone.env with a local or remote OpenAI-compatible endpoint.

3. Add Voice Use windowed setup when needed, paste an ElevenLabs API key, load voices, choose one, and test speech.

4. Set Presence Choose the wake word, screen timeout, memory behavior, and known users in Companion settings.

5. Launch Kiosk Run ~/synra-jetson-station/scripts/start-electron-kiosk.sh for the verified 30 FPS path.

Read Install Guide Operations Runbook