A voice cursor for your computer. Hold a key, talk into any text field — email, code editor, chat, terminal. Speechcap transcribes, cleans, and pastes the result where you were already typing.
No window to open, no recordings to transcribe later. Push-to-talk, like a walkie-talkie for writing.
Right Option by default — or fn, ⌘, or Caps Lock. Whatever you'll never press by accident.
Filler words, restarts, second thoughts — all fine. Speechcap cleans them up.
Text is pasted right where your cursor was. No window-switching, no copy-paste dance.
Select text and press your hotkey to open the transform menu — or hold a letter while dictating to apply one on-the-fly.
Highlight a paragraph, tap your hotkey, pick the transform. Reshape what's already on screen.
Settings stay flat — three groups per screen, no nested submenus. History lives one click away. Billing is honest about what you're using.
Four screens. Permissions, hotkey, a guided test, then you're typing with your voice.
Free is meaningfully useful — 2,000 words a month, raw transcription. Pro unlocks the rest.
On Pro, yes — on-device Whisper transcribes without a network. The Free tier and Pro cloud mode both need an internet connection.
Free and Pro cloud mode send audio to our transcription provider over an encrypted connection — it's discarded immediately after transcription. On Pro on-device mode, audio never leaves your Mac.
Audio is encrypted in transit and at rest. We never train on your data. We don't share with third parties.
Yes — toggle "tap to start" in Settings. We default to push-to-talk because it's less ambiguous about when we're listening.
Mac only today (Apple Silicon and Intel). Windows is next on the roadmap; Linux after that.