Voice input
Dictate messages into any chat, choose which microphone Slashspace listens to, and fix dictation when it doesn't work.
Thinking out loud is often faster than typing, especially when you're reacting to a source or talking through a half-formed idea. Voice input turns what you say into text in the message box, so you can ramble first and tidy up before you send.
Dictation works in every chat composer: in a Chat node on the canvas and in the side view. It isn't available in Text nodes.
Dictate a message
- Click the microphone button in the composer toolbar, next to the model button. Or use a shortcut from the table below.
- Speak. A timer and a live waveform show while Slashspace listens.
- Press
Enteror click the check mark. After a moment of "Transcribing...", the text appears in the message box. - Edit the text if you need to, then send it as usual. Nothing reaches the AI until you send it.
| Action | Shortcut |
|---|---|
| Start dictating (message box empty) | Space twice, quickly |
| Start dictating (cursor in the message box) | Ctrl+M (the Control key on a Mac too, not Command) |
| Finish and transcribe | Enter |
| Cancel and discard the recording | Esc |
Note: The transcript replaces whatever is already in the message box, so dictate first and type afterward.
Choosing a microphone
Slashspace uses your system's default input unless you pick another. Two places set the same preference, and it applies to every chat:
- From the composer: hover the microphone button and click the small arrow beside it. In the Microphone menu, pick System Default or a specific device.
- From Settings: open Settings > System Preferences and choose a Preferred Microphone under Audio Settings.
If the microphone you picked is unplugged, dictation won't start until you switch back to System Default.
How transcription works
Transcription uses Groq's Whisper speech-to-text model, so it needs an internet connection. The recording is sent only when you finish it; a cancelled recording never leaves your computer.
- Without a Groq key (the default), your audio goes through Slashspace's server to Groq. There's nothing to set up, and it doesn't draw from your hosted wallet.
- With your own Groq key saved in API providers, Slashspace sends the audio straight to Groq, and the usage counts against your Groq account.
See Privacy and your data for what else leaves your computer.
Permission and troubleshooting
On a Mac, the first time you dictate (or open the microphone list), macOS asks whether Slashspace can use the microphone. Allow it.
If the recording closes right away, or no text appears after "Transcribing...", Slashspace doesn't show an error. Check:
- Mac: System Settings > Privacy & Security > Microphone has Slashspace turned on. If you declined the prompt earlier, macOS won't ask again.
- Windows: your privacy settings let desktop apps use the microphone.
- Microphone: System Default is selected in the Microphone menu.
- Internet: you're online.
- Groq key: if you added one, it's still valid in API providers. Removing it switches back to the built-in transcription.
Still stuck? See Troubleshooting and support.