Voice dictation for Pi.

Pi’s core has no voice input. Voice comes only from community pi packages such as pi-voice-stt or picrophone, which need a cloud speech API key (OpenAI, Groq, Deepgram, Mistral) or a local speech model you set up yourself. pi.dev reminds you that third-party packages run code on your machine and should be reviewed before installing.

As of October 2026.

• No pi package to install, review or keep up to date: Copy That pastes into the terminal pi runs in.
• Transcribed on your Mac with Whisper, so no audio goes to a speech service and there’s no speech API key to set up.
• One button on your phone or watch instead of a hotkey, and you can talk from across the room.
• Works the same with every model and provider you use in pi.

1. Push.

   One button on your Apple Watch or iPhone.

2. Talk.

   Say the prompt, anywhere, on any network.

3. Paste.

   Your Mac transcribes it and pastes it into Pi.

Does Copy That work with Pi?

Yes. Copy That pastes your transcribed words into whatever app is in front on your Mac, so it works with Pi with no plugin or setup inside Pi.

Does pi have voice input?

Not built in. Voice comes from community pi packages such as pi-voice-stt or picrophone, and most of them need a speech API key or a local model setup.

Do I need a pi extension to dictate with Copy That?

No. Copy That transcribes on your Mac and pastes the text into the terminal where pi is running, so pi itself needs nothing installed.

Do my iPhone and Mac need to be on the same Wi-Fi?

No. Recordings go through your own private iCloud, so your phone can be on cellular in a café while your Mac is at home. The Mac just needs to be awake and online.

Where is my voice transcribed?

On your Mac, with Whisper. Audio passes through your own iCloud account and is never sent to a speech-to-text service.

What does it cost?

The iPhone and Mac apps are free. The Apple Watch app is a one-time $4.99. There is no subscription.