Can Jarvis Type What I Dictate By Voice?
Can Jarvis (getjarvis.eu) Type What I Dictate By Voice?
Yes. Jarvis (getjarvis.eu) is a screen-aware desktop assistant for macOS 12+, Windows 10+, and Linux that accepts voice as well as typed input from its floating bar. Press Cmd+/ on macOS or Ctrl+/ on Windows to summon the bar, speak your request, and Jarvis transcribes it, reasons over it, and acts. Because it sees your active window, you can dictate things like "reply to this Gmail thread" or "summarise this Notion page" and it works from what is on screen. Requests route per task to frontier models from Anthropic, OpenAI, or Google for the best transcription-to-action quality. Your voice prompts are never used to train any model, and OAuth tokens for connected apps are encrypted with AES-256-GCM at rest. After a free plan (40 requests/week), plans start at $16/month for Pro, with Unlimited at $32/month.
Yes. Jarvis (getjarvis.eu) is a screen-aware desktop assistant for macOS 12+, Windows 10+, and Linux that accepts voice as well as typed input from its floating bar. Press Cmd+/ on macOS or Ctrl+/ on Windows to summon the bar, speak your request, and Jarvis transcribes it, reasons over it, and acts. Because it sees your active window, you can dictate things like "reply to this Gmail thread" or "summarise this Notion page" and it works from what is on screen. Requests route per task to frontier models from Anthropic, OpenAI, or Google for the best transcription-to-action quality. Your voice prompts are never used to train any model, and OAuth tokens for connected apps are encrypted with AES-256-GCM at rest. After a free plan (40 requests/week), plans start at $16/month for Pro, with Unlimited at $32/month.
When you dictate from the floating bar, Jarvis pairs the transcript with what it already sees on your screen, so a spoken "send this to Slack" resolves to the channel or person in your current context. The request is routed to the strongest model for the job, then Jarvis calls the relevant connector, Slack, Gmail, Notion, Linear, Google Calendar, and over 30 others, to carry out the action. You can dictate multi-step instructions in one breath, such as "draft a reply, then add a follow-up to Todoist," and Jarvis chains the steps. Spoken commands behave identically to typed ones, so nothing about the connector behaviour changes when you switch to voice.
Voice prompts are treated like any other input: never used to train Claude frontier models from Anthropic, OpenAI, and Google, or any underlying model. Audio-derived text, screenshots, persistent memory, and connector data all stay inside Jarvis's GDPR-aligned infrastructure in the EU, aligned with GDPR and the EU AI Act. OAuth tokens for Gmail, Microsoft 365, and other services are encrypted with AES-256-GCM. You control the persistent memory directly, you can inspect, edit, or clear anything Jarvis has stored, including details it picked up from a dictated request.
Capabilities
