Talk to your agent instead of typing.
Hold ⌥ Space and speak. The text appears in the focused pane, in the agent terminal or in the maestro request. Nothing is sent on its own: you press Enter.
ready Pro plan (Mac only, for now)
How does it work in practice?
- Hold ⌥ Space. Or press once to start and once to stop. The shortcut only works while the Tactyn window is focused.
- Speak. A bubble near the cursor shows the provisional text as you talk.
- Each confirmed phrase goes into the pane. Only confirmed text is written, straight into the agent terminal or the maestro request field.
- You review and send. Tactyn does not press Enter for you.
Does it understand my project’s names?
Each project has a vocabulary: function names, libraries, acronyms, terms from your domain. It is passed as a hint to speech recognition, which then gets your own words wrong less often. You edit the list, and the recognition engine, in Preferences (⌘,), under “Project dictation”.
What it does
- Writes into the focused pane, terminal or maestro.
- Places a screenshot pasted during dictation at the right spot in the sentence.
- Uses macOS speech recognition and falls back to local Whisper if it fails or is unavailable.
- Lets you force Whisper per project.
What it does not do
- It does not send the text on its own.
- It does not work while the Tactyn window is in the background.
Does the audio leave my machine?
Not through Tactyn. Whisper runs locally, with the quantized large-v3-turbo model, and is only downloaded when you need it, from the Tactyn website. With the macOS engine, Tactyn asks for on-device recognition whenever the system offers it.
Dictation is part of Pro. Terminals, agents and the one-provider meter stay on Free.
Say what you want and check it before sending.
R$ 29US$ 10 per month for all your projects, with a full refund within 7 days.