This is a read-only demo of the Settings page. In the real product, all settings are fully configurable.

Settings

System configuration, security, and portability

AI Model
Active Modelgemma-3-27b-it
ProviderSpiritBox LLM Engine
Context Window131,072 tokens
Auto-load on startup
Temperature0.7
Voice Settings
TTS EnginePiper TTS (Local)
Wake Wordhey spirit
Voice Input (STT)
Voice Output (TTS)
Iron Dome (Feedback Prevention)
Security
Egress Firewall
Aegis Sanitizer
Data Sanitizer
Credential VaultAES-128 Encrypted
Outbound TelemetryDisabled (always)
Connectivity
Telegram Bot
Remote AccessTelegram Gateway
Web SearchSearXNG (Self-hosted)
n8n Workflows12 Active
Memory System
Knowledge GraphNeo4j (42 nodes)
Vector Storepgvector (128 embeddings)
Human Approval Gate
Auto-Detect Facts
Session Buffer3 pending
Portability

SpiritBox Ultimate is a physical asset. Your entire AI ecosystem — conversations, memories, Spirit Guides, voice models, and settings — lives in one directory. Copy it to any machine with Docker and you're running in minutes.

Total Data Size2.4 GB
Last Backup2 hours ago
What SpiritBox is NOT
NOT a cloud service
Everything runs on hardware you own. No monthly fees. No vendor lock-in. No "we're shutting down" emails.
NOT always listening
Zero passive monitoring. Zero telemetry. The system only activates when you trigger it. Period.
NOT a web scraper
Web access goes through isolated "air-locks" (SearXNG container). The AI never browses freely.
NOT training on your data
Your conversations, creations, and knowledge are never used to train models. They're yours. Full stop.
Back to Overview