Voice
Speech to text types what you say, with a model on your Mac. Voice Control talks with you about your workspace, on a cloud model.
Speech to text
Speech to text works in the desktop app on a Mac. It dictates into a chat, a note, a field or a terminal. Recognition runs on your Mac; Ruimte never uploads or saves audio, and you need no account or API key.
Turn it on
- Open Settings, Voice.
- Turn on Enable Speech to text. Ruimte downloads a 2.6 GB speech model the first time.
- Pick the microphone and the language.
Turning Speech to text off keeps the model. Remove model deletes it.
Dictate
Press CtrlShiftD to start and to stop. A live transcript shows while you speak. In a terminal, the text waits until you click Paste into terminal; it never presses Enter.
Voice Control
Ask Voice Control what is on screen, switch to another view, add a note, open a terminal, start a chat or send a chat a message.
Turn it on
- Open Settings, Voice, Voice Control.
- Enter your OpenAI API key. It stays on your machine and goes only to OpenAI.
- Pick its voice and the language it speaks.
Talk
Press CtrlShiftM to start or end a conversation, or hold it to talk. The Voice Control panel shows what you said, what it answered and each action it took. Undo action reverses a note it added.
| Limit | Detail |
|---|---|
| Project | It works in the current project of one machine. |
| Chats it starts | Supervised mode at most |
| Clearing a chat | It asks you first, every time. |
| Deleting views or nodes | It asks first while Confirm deletions is on. |
| Audio | The machine never keeps it. |