About
Whisperstream is a Windows-native dictation app that runs 100% on your PC. Press a push-to-talk hotkey, speak, and your transcribed words paste into whatever window has focus: your IDE, your email, your notes, Slack. Transcription runs locally on the CPU via NVIDIA Parakeet or Qwen3 ASR. Audio never touches the network.
Key features:
Private: 100% local, on-CPU transcription. No network calls during dictation.
No Subscription: $29 one-time purchase.
Fully Local or Cloud Enhancement: Custom prompts or basic cleanup
Automatic Profile Switching: Customize enhancements for each app or website
Transcript History: Store and search through past transcripts
File Input: Transcribe your audio files
Multilingual: Supports 39 languages including English, European, East Asian and South Asian
Push-to-talk hotkey or toggle mode. Bind to any key.
Auto-lowers your music while you dictate.
Free trial on first install, no account required.
Built for Windows vibe coders dictating into Cursor, Claude Code, or VS Code; privacy-critical professionals (legal, healthcare, defense); accessibility and RSI users; writers; and academics. The wedge: Windows-native, fully local, one-time pricing. Differentiation versus cloud-tied dictation tools that charge a subscription or fail compliance requirements.