Documentation
Personal Jarvis docs
Everything from a one-line install to the architecture behind the voice. New here? Start with the introduction, then install it.
Getting Started
- Introduction What Personal Jarvis is, what it is not, and the one idea that makes it different.
- Installation Installation paths for Windows, macOS, and Linux. Windows is the hardware-tested desktop target.
- First Run & Your Keys The setup wizard, where your API keys actually live, and how to talk to Jarvis the first time.
Core Concepts
- Architecture The eight-layer model, the strict dependency rule, and why every backend is a plugin.
- Voice Pipeline From wake word to spoken answer — the path every utterance takes, and the sub-second Ack that hides the latency.
- Brain Providers Five interchangeable model providers, switching by voice, and the fallback chain that keeps you running.
- Harness Dispatch How the Router-Brain decides where a command goes, and what a harness actually is.
Guides
- Missions & the Critic Loop How complex tasks run in isolation, get reviewed, and heal themselves before Jarvis speaks back.
- Computer-Use & Risk Tiers Letting Jarvis drive your screen safely — the four-tier risk system that fights confirmation fatigue.
- Configuration jarvis.toml, atomic config writes, and how Jarvis safely rewrites its own settings at runtime.