Hold one key and talk. LiveDictate transcribes your speech, cleans it up with AI, and types it straight into whatever app you're using.
Everything you need to replace the keyboard for the messy first draft, the quick reply, and the long email you keep putting off.
Hold your hotkey to dictate and release to insert. Or press once to start and again to finish. Your keys, your call.
CtrlYour words land wherever the cursor is. No copy and paste, no separate window to babysit.
Speak naturally in your language. LiveDictate detects it automatically, no settings to fiddle with.
Music and video mute while you dictate, then pick right back up when you finish.
Add names, jargon, and product terms to a custom vocabulary so they're spelled right every time.
Download for Windows and run it. Pick a hotkey in the quick setup, and you're live.
Click into any text field, hold your key, and talk. A small overlay shows it's listening.
Let go. Clean, formatted text drops in at your cursor in about a second.
Two transcription models, one toggle. Trade a beat of latency for accuracy whenever the words have to be exactly right.
Turn on Smart mode and an AI model cleans up each take as you dictate. It fixes the filler, the false starts, and the punctuation so you don't have to.
Faithful, casual, professional, or concise. Match the tone to the moment.
Roughly a second of polish before the text lands, no second pass needed.
um so yeah can you like send over the uh the q3 numbers when you get a sec thanks
Could you send over the Q3 numbers when you get a chance? Thanks.
LiveDictate doubles as a control room for Claude Code. Drive several projects at once, hear what each one is doing, and approve commands by voice from a floating agent bar.
Every session shows its branch, last message, and progress at a glance.
Say approve or deny when an agent asks to run a command. No tab switching.
LiveDictate only captures audio while you hold your key — never in the background. Your speech is sent over an encrypted connection to best-in-class transcription models, turned into text, and never stored by the app.
Give your hands a break and let your voice do the writing. Set up takes about a minute.