Voice-to-Text for Windows

Record speech from your microphone, transcribe it with OpenAI Whisper, and clean it up with AI — built specifically for Windows.

Features

• Real-time audio visualization — see your microphone waveform while recording
• Multiple transcription models — choose between GPT-4o Transcribe (best quality, $0.006/min), GPT-4o Mini Transcribe (good quality, $0.003/min), or Whisper v2 (legacy, $0.006/min)
• 50+ languages — auto-detect or select from all Whisper-supported languages
• AI cleanup modes — General, Email, Slack/Chat, Notes, or your own custom prompt
• Dictation history — browse, search, copy, and paste past transcriptions
• Auto-copy to clipboard — results are copied automatically after transcription

Setup

1. Get an OpenAI API key here: https://platform.openai.com/api-keys
2. Open extension preferences and paste your API key
3. Optionally configure language, transcription model, and prompt mode

Commands

Dictate

Record speech and convert it to clean text. Press Enter to stop recording and begin transcription. The result is automatically copied to your clipboard.

Dictation History

Browse and search past dictations. Copy or paste previous results directly from the list.

Select Mode

Quickly switch between AI cleanup modes:

• General — Fix grammar, punctuation, and filler words
• Email — Format as a professional email with subject line and sign-off
• Slack / Chat — Casual, concise messaging tone
• Notes — Structured bullet points
• Custom — Use your own prompt (configured in preferences)

Requirements

• Windows 10 or later
• A working microphone (uses your system default from Windows Sound settings)
• OpenAI API key (for transcription and AI cleanup)
• Raycast Pro is optional - if you have it, AI cleanup uses your selected Raycast AI model. Otherwise falls back to gpt-4.1-nano.