OpenWhispr turns a hotkey into free voice dictation that types at your cursor on Mac, Windows, and Linux
The fastest way to write something is often to stop typing it.
OpenWhispr is an MIT-licensed dictation app that states its whole job in one line: press a hotkey, speak, and your words appear at your cursor. It ships finished installers rather than a developer checkout, a .dmg for macOS on both Apple Silicon and Intel, an .exe for Windows, and AppImage, .deb, .rpm, and .tar.gz builds for Linux, and the project says every core feature, transcription, AI reasoning, speaker identification, and semantic search, will run on local models with GPU acceleration through Metal, CUDA, or Vulkan, with cloud providers offered as a choice rather than a requirement. Past plain dictation it will transcribe an audio file or a recorded meeting and label who was speaking. The one stated limit belongs to Intel Macs, where live speaker identification and voice fingerprinting are unavailable because ONNX Runtime no longer ships x86_64 binaries, though meetings still record and semantic search falls back to keyword matching.
