Draft at the speed of thought
Long emails, notes, that first messy paragraph - say it out loud and it lands as text, far faster than you can pick it out key by key.
Press one hotkey, talk, and ClintSpeaks types what you say straight into whatever app has focus - a browser, an email, a doc, even your code. Fully on your machine.
Honestly, the fastest way to get a thought down isn't to type it out, it's to just say it and watch the words land
Most dictation is buried in OS settings or locked to one app. ClintSpeaks just types what you say - anywhere - and gets out of the way.
Long emails, notes, that first messy paragraph - say it out loud and it lands as text, far faster than you can pick it out key by key.
Sore hands, RSI, a wrist that flares up by mid-afternoon? Give your hands a break and keep working with your voice instead.
Pacing, sketching, hands busy elsewhere? Talk and the thought is captured before it slips away - no keyboard needed.
Keep your eyes on the work, not the keyboard. Speak a sentence, glance up, it's already there - fewer context switches, less friction.
No setup wizard, no bloat, no account. Install it, learn one hotkey, and it's part of how you work.
Put your cursor in any app, press Ctrl+Shift+D, and talk. The words are typed in for you - exactly where you were already working.
Pick what fits: push-to-talk (hold the key while you speak), tap-to-toggle (tap on, tap off), or auto-stop that ends when you go quiet.
A focus check makes sure the text goes into the window you were actually typing in - never the wrong app, never a surprise.
Match your desktop automatically or pick a mood. The floating mic HUD and settings look right against any background.
Teach it the names, jargon and acronyms you actually use, so they come out spelled right the first time - not phonetic guesswork.
Lives quietly in your system tray. No taskbar button, not in Alt-Tab - there when you need it, invisible when you don't.
It backs up your clipboard before inserting text and restores it after. Whatever you had copied stays exactly where you left it.
Flip one toggle and it's ready the moment you sign in. Tiny footprint, instant launch, no waiting around.
Start dictating the moment it's installed with a built-in on-device model, or switch on a faster or more accurate one. Every engine runs locally.
An accurate, multilingual speech model that runs fully offline - no cloud, no account, nothing leaves your machine. MIT-licensed and free. To keep the installer tiny it downloads once on first run, then it's always ready. Got an NVIDIA card? Optional GPU acceleration transcribes long dictations in a blink. Prefer the fastest possible response? Switch to the tiny Moonshine engine with a click.
Pick the engine that fits the moment - fastest, most accurate, or most languages. One click to switch, all on-device.
Everything happens on your computer. The model runs locally, your speech is transcribed on-device, and the text goes straight to your cursor. No cloud, no account, no upload - so even your most sensitive notes, messages and ideas stay yours. The easiest way to trust it is the simplest test: pull your network cable and keep dictating.
Honest note: this isn't a medical, clinical or legal transcription tool, and accuracy varies with your accent, your microphone and background noise. ClintSpeaks is built for everyday writing by voice - fast, private, and out of your way.
ClintSpeaks is a complement, not a replacement. If you need to drive your whole computer by voice - every menu, button, click and command - these are the leading tools, and they're excellent. Use them, and keep ClintSpeaks alongside for the moments you just want to talk and have it typed.
No account, no trial, no sign-up. Install it and use it as much as you like - if it earns a place in your day, a one-time coffee says thanks.
Built by one person who'd genuinely rather talk than type. A one-time coffee (from $2) says thanks - and switches off the occasional in-app reminder.
Support developmentOptional. No spam - just the occasional note when something good ships.