OpenWhispr/openwhispr
Desktop voice-to-text app that turns speech into text at the cursor through a global hotkey, with optional AI commands, meeting transcription and notes. Transcription can run fully offline with Whisper, NVIDIA Parakeet or Cohere models, or in the cloud, and speaker diarization and semantic search also work locally. Built with Electron, React and TypeScript, released for macOS, Windows and Linux under MIT.
Maintenance activity