Voice is faster than typing. Most of us talk at well over a hundred words a minute and type at a fraction of that; in a Stanford study, speech input was about three times faster than typing on a phone keyboard. Dictation should be the easy win.
But the most popular dictation tools send your voice to a server to be transcribed, and charge you a monthly fee for it. Your voice is personal. You say things out loud that you would never write in an email. You shouldn't have to rent it back from someone else's cloud.
You don't have to anymore. Open speech models like Parakeet and Whisper got good enough, and Apple Silicon got fast enough, that a Mac can recognize speech on its own and keep up with you. So that is what Hush does.
Free
Hush is free. Not for 14 days, not a free tier with a word limit, not “free while in beta”. There is nothing to buy and no account to create. It costs almost nothing to run software on your own computer, so it shouldn't cost you $15 a month.
Private
Your audio is recognized on your Mac and is never uploaded. Hush has no analytics and no telemetry. The only time anything leaves your Mac is when you ask it to: downloading a speech model, or turning on AI clean-up with a provider you chose, in which case the text (not the audio) goes to that provider under your own key. We wrote down exactly what leaves your Mac.
Open
“Trust us” is not a privacy policy. Hush is open source under the MIT license, so you can read every line that touches your voice, build it yourself, and change it to suit you. If Hush ever stopped being what it says it is, anyone could fork it.
What Hush is not
It is not a meeting recorder and it doesn't watch your screen. It does one thing: you hold a key, you talk, and your words appear where your cursor is.
Deepanshu Pal
Maker of Hush