Prerequisites
Prepare the runtime, accounts and devices for the current voice-first source package.

Python and uv
The current core package declares Python 3.10 or newer. Use uv to create the environment and install the locked dependencies. The assistant no longer needs a Bun or frontend build; the website in the Frontend folder is a separate application.
uv --versionConfigure both brains
You need Gemini Live credentials for voice and a configured task provider for delegated work. Set GEMINI_API_KEY in the project-local .env. Configure the high-level provider with jarvis model and supply the credentials it requires.
Cloud providers and external services may charge independently. Source access does not include model API usage.
Devices and desktop capabilities
Live audio needs a working microphone, speaker and voice dependencies. The detailed native desktop, Explorer and local OCR workflows described here target Windows. Other environments can have different available capabilities.
Browser organization needs the connected Chromium extension; editor workflows need the VS Code adapter. Video/audio preparation needs FFmpeg and ffprobe. Inspect status on the actual machine rather than assuming an integration is installed.
uv run jarvis desktop status
uv run jarvis lowlevel --list-devices