Getting started
From a new account to your first spoken answer.
1. Create your Thundera account
Go to auth.thundera.eu/register. One account signs you in to Jarvis, EAM and every other Thundera app.
2. Install the desktop app
The desktop app is macOS (Apple Silicon). The signed release build is being prepared; the landing page shows where to get it when it is published. Windows and Linux are not available.
The first time you use voice, Jarvis downloads its local speech model once in the background (sha256-verified).
3. Sign in
Open Jarvis and choose Sign in. It opens your browser for the Thundera login and brings you back. The same sign-in connects EAM, so your memory works straight away.
4. Give Jarvis a brain
| Option | How | Cost |
|---|---|---|
| Your own key | Settings, then paste a DeepSeek, OpenAI, Anthropic or OpenRouter key | The provider bills you directly |
| Hosted models | Subscribe to Basic or Advanced | Included in your plan allowance |
Your own keys are stored in macOS secure storage. See Bring your own key.
5. Talk
- Manual mode (default): click Speak, talk, and stop. Recording ends after 5 seconds of silence. Hold the Space bar for push-to-talk.
- Active listening: switch it on (⌘⇧L). Say "Hey Jarvis" or just "Jarvis". See Voice for the one-time wake model install.
- Typing: press ⌘L and write to Jarvis.
Press ⌘K for the command palette: everything Jarvis can do is there.
Keyboard shortcuts
| Shortcut | Action |
|---|---|
| ⌘K | Command palette |
| Space (hold) | Talk |
| Esc | Stop or close |
| ⌘1 to ⌘9, ⌘0 | Sections (⌘0 is Actions) |
| ⌘, | Settings |
| ⌘⇧M | Mini mode |
| ⌘⇧L | Active listening |
| ⌘⇧G | Camera and gestures |
| ⌘⇧K | Stop screen control |
| ⌘⌥⇧P | Privacy mode (works from any app) |