Download

Everything you need. Nothing you don't.

WhisperType is built for people who want fast, private, local voice dictation on Windows or Mac.

Download WhisperType Pricing — $9/mo

Core features

  • 100% local transcription

    Your voice stays on your machine. No audio sent to the cloud, no transcripts stored on servers. Processed locally by Whisper models running on your own hardware.

  • GPU acceleration

    WhisperType uses your NVIDIA GPU through CUDA on Windows and Metal on Apple Silicon, automatically. There is nothing to switch on.

  • Language detection

    Auto chooses between English and Spanish. For another language, pick one of twelve in Settings.

  • CPU fallback mode

    No GPU? WhisperType runs on the CPU. Slower, same features.

  • Works offline

    After the one-time model download, transcription needs no internet connection. The app still goes online for license checks, updates and anonymous statistics; the statistics can be turned off.

  • Native on Windows and macOS

    A native app on each: in the system tray on Windows, in the menu bar on macOS.

Workflow features

  • F8 hotkey

    Default F8 hotkey starts/stops recording. Change it in Settings → Hotkey. It works whichever app is in front.

  • Floating glass widget

    A small floating bar shows the recording status and a waveform. Drag it anywhere on screen.

  • Waveform visualization

    Real-time audio waveform display helps you see when you're speaking and monitor audio levels.

  • System tray presence

    Minimizes to system tray. Right-click for quick access to Settings, License, and Exit.

  • Launch at login

    Option to launch WhisperType when you log in. Always ready when you need it.

  • Works in any app

    Dictate into Discord, Slack, VS Code, Cursor, Chrome, Firefox, Word, Notepad, terminal — any app with text input.

Built for developers & agentic workflows

Use WhisperType while running Cursor, Claude Code, Codex, or any agentic tool. Don't break your flow to type context. Press F8, describe the change, press F8. Done.

"Give your agents more context by talking through your change instead of typing it."

Learn About Agentic Workflows

AI Polish & productivity

  • AI Polish (opt-in)

    Optional post-transcription cleanup via local Ollama (default, llama3.2:3b) or OpenAI (opt-in). 2-second timeout cap; falls through to the raw transcript on any error. Your voice never leaves the machine; only the text is sent to the provider you chose.

  • Custom dictionary

    Add proper nouns, brand names, and technical terms so Whisper gets them right every time. Your words are given to Whisper as a hint on every transcription. Nothing is retrained and nothing is uploaded.

  • Text snippets

    Map short triggers to longer expansions applied after transcription — signatures, boilerplate, code templates. Dictate the trigger, get the full block.

  • Transcription history

    Searchable local log of everything you've dictated. Stays on-disk, never synced anywhere. Copy or delete any entry.

  • Settings panel

    Change hotkey, language, model, startup behavior, polish provider, dictionary, and snippets from a proper UI. No config-file editing.

  • First-run onboarding

    Setup walks you through the microphone, the hotkey and a first dictation. Anonymous usage statistics are on by default and can be turned off; crash reports stay off until you turn them on.

Reliability

  • Hot-plug recovery

    Re-initialises the audio subsystem when your default input device disappears — unplugged headset, Bluetooth disconnect, USB re-enumeration — so dictation keeps working without a restart.

  • Self-healing watchdog

    A per-user scheduled task checks every 5 minutes that the app is alive and relaunches it if not. Registered by the installer, removed on uninstall — clean box.

  • Single-instance lock

    Launching a second copy exits cleanly. No duplicate panels, no duplicate tray icons, no fighting hotkeys.

  • In-app update banner

    Checks whispertype.ai for a newer version 30 seconds after launch and every six hours. On Windows the update downloads in the background and installs when you press Restart to install, or the next time you quit. On macOS the banner opens the download and you install it yourself.

  • Opt-in crash reports

    Disabled by default. When enabled: only stack traces + non-personal diagnostic tags (OS, app version). Before-send scrubbing guarantees no audio, transcribed text, or clipboard content ever leaves the machine.

What's next

Roadmap items currently in development.

  • Signed installer

    Code signing on Windows and macOS, so the first launch needs no "Run anyway" or "Open Anyway".

  • Coding mode

    Code-aware punctuation, PascalCase / snake_case handling, VS Code + Cursor integration hints.

See full changelog