Your voice becomes your keyboard
Uptype turns your voice into text anywhere on macOS and Windows. Dictate text or give an instruction: Uptype writes the message, email, or reply ready to send.
You spend your day typing.
Your fingers are 4× too slow.
Hold
⌥ · speak · it’s typed.
One key, your voice, your text. Uptype listens locally, transcribes with sharp punctuation, and adapts the format to the app you're in: Slack, Notion, Gmail, VS Code, your terminal, anywhere.
- ✓Latency < 400ms. you finish speaking and the text is already there
- ✓Context-aware format. line breaks in an email, short message in a chat, bullet lists in a document
- ✓Auto punctuation. commas, periods, line breaks, without dictating them
- ✓Multilingual. French, English, Spanish, German, Italian
- ✓Everywhere you type. native fields, web apps, IDEs, terminal
Voice command
Say the intent.
Uptype writes the text.
Turn on voice commands and ask Uptype to generate an email, reply, message, or note from scratch. If you dictate, Uptype transcribes. If you give an instruction, Uptype produces the final text.
Auto-detection
The right format
in the right app.
Uptype detects which app you're writing in and adapts the formatting automatically. No more reformatting after dictation — line breaks, tone, and structure are ready to go.
Hi Mark,
i think this is a good idea but we should discuss it monday because i have not had time to review it in detail
sent from Uptype
Highlight. Speak.
No more copy-paste.
On top of auto-formatting, select any text. Say an instruction out loud. Uptype rewrites it in place — shorten, translate, fix, without opening a tab.
Configured translation
Speak French.
Write English.
Choose a target language in Uptype once. Then dictate normally: the text is transcribed, translated, and inserted into the open app with the right formatting.
- ✓Set once. the target language stays active until you change it
- ✓Still contextual. structured email, short message, or note depending on the app
- ✓No selection needed. unlike editing, translation happens while you dictate
Same sentence. Two worlds.
Simple. Cancel anytime.
Pro monthly
For professionals who type all day.
- ✓7-day free trial
- ✓Unlimited dictation
- ✓Context-aware auto formatting
- ✓Configured translation
- ✓Voice generation and editing
- ✓Custom templates
- ✓Business vocabulary
- ✓Priority support
Pro yearly
The same Pro access, with two months free.
- ✓7-day free trial
- ✓Unlimited dictation
- ✓Context-aware auto formatting
- ✓Configured translation
- ✓Voice generation and editing
- ✓Custom templates
- ✓Business vocabulary
- ✓Priority support
Transparency
What Uptype does on your computer
and what it never does.
Uptype is a push-to-talk dictation tool. To turn your voice into text and paste it where you are typing, it needs a few system-level capabilities. We list every one of them below in plain language so you know exactly what is happening on your machine. The desktop binary is signed by SAS VRM using Microsoft Azure Trusted Signing on Windows and Apple Developer ID notarization on macOS.
Records audio only while you hold the push-to-talk shortcut
The microphone is closed by default. It opens the moment you press your configured shortcut and closes the moment you release it. Audio is sent to our transcription gateway over HTTPS, transcribed, and discarded. No background recording, no continuous listening.
Listens for one specific key combination, nothing else
Uptype uses Apple's documented event tap API, gated by the macOS "Input Monitoring" permission you explicitly grant. It only inspects whether the configured shortcut (e.g. Globe/Fn) is pressed. It does not read, store, log, or transmit any other keystroke. The listener stays off until you complete the onboarding step that enables it.
Detects the configured key, lets every other key pass through
Uptype uses Windows' low-level keyboard hook API solely to detect the configured push-to-talk key (e.g. right Ctrl). All other keys pass through unchanged. The hook never records, saves, or transmits anything beyond "shortcut pressed" / "shortcut released" signals.
Reads the text you have selected, only when you dictate
If you have text selected when you dictate, Uptype uses macOS Accessibility APIs to read that selection so it can rewrite it according to your instructions. This requires the "Accessibility" permission you explicitly grant in System Settings. Without selection, nothing is read.
Pastes the result into the field you are typing in
After transcription, Uptype temporarily writes the result to the clipboard, simulates a Cmd+V (macOS) or Ctrl+V (Windows) to paste it, and then restores your previous clipboard content. The simulated paste is the only synthetic keystroke Uptype ever sends.
All transcription happens on our signed servers
Your recording is uploaded over HTTPS to api.uptype.ai. Transcription happens server-side. Nothing is shared with third parties beyond the speech-to-text provider, which is bound by our data processing agreement. The Privacy Policy details every endpoint.
The Windows installer, the main Windows executable and every bundled binary are signed by SAS VRM with Microsoft Azure Trusted Signing (Microsoft-attested identity verification). The macOS installer package is signed and notarized by Apple. The signatures are verified automatically by Windows SmartScreen, Apple Gatekeeper and your operating system before the app ever runs.
Every system permission Uptype requests is granted by you through your operating system's native dialog. You can revoke any of them at any time from System Settings (macOS) or Settings → Privacy & security (Windows), and Uptype will keep working with the remaining features. Uninstalling Uptype removes the app and the bundled helper from your machine.
The desktop app, the website, the privacy policy and the API are all documented publicly. You or any reviewer can audit exactly what is in the binary you install.