Voice Dictation

Voice Dictation lets you speak into your microphone and have the audio automatically transcribed by Gemini AI. The transcribed text is inserted directly into the current Cursor/VS Code chat input, GitHub issue description, or issue comment — no manual copying required.

Caution

Voice Dictation is currently Windows only. It uses the built-in winmm.dll MCI API — no additional software to install.

Requirements

• Windows (built-in winmm.dll — nothing to install)
• Gemini API key configured in settings (devManager.gemini.apiKey)

Get a free key at aistudio.google.com.

---

Status Bar Button

A microphone button 🎙 appears in the status bar at the bottom of the IDE (Windows only).

1. Click 🎙 to start recording. The icon turns red ⏹.
2. Speak your message.
3. Click the red ⏹ to stop. The icon shows a spinner while Gemini transcribes.
4. The transcribed text is pasted automatically into the current chat input.

The chat panel is focused automatically — you can keep dictating into the same conversation as many times as you want without creating a new chat.

Keyboard shortcuts

Shortcut	Action
Ctrl+Alt+H	Toggle recording (start / stop)
Ctrl+Alt+Shift+H	Cancel recording

These match package.json → contributes.keybindings: both are bound with when: "!terminalFocus" so they do not fire while the integrated terminal has focus (avoid stealing keys from the shell).

---

Voice Dictation in GitHub Issues

Microphone buttons are also available directly in the New Issue and Edit Issue forms.

Description field

Click 🎙 next to the Description label → speak → click again to stop. The transcribed text is appended to the description textarea.

Comment field (Edit Issue only)

Click 🎙 next to Add comment → speak → stop. Text is appended to the comment textarea.

---

Language Setting

By default, Gemini auto-detects the spoken language. To force a specific language:

"devManager.gemini.dictationLanguage": "uk"

Value	Language
auto	Auto-detect (default)
uk	Ukrainian
en	English
de	German
fr	French
es	Spanish
pl	Polish
it	Italian
pt	Portuguese
nl	Dutch
cs	Czech
ja	Japanese
zh	Chinese
ko	Korean

---

Cost

Voice dictation uses the same Gemini API key as title generation. Pricing for Gemini 2.5 Flash:

Duration	Cost
30 seconds	~$0.00009
5 minutes	~$0.0009
1 hour of dictation	~$0.011

At typical usage (a few minutes per day), the cost is well under $0.05/month.

---

Settings Reference

Setting	Default	Description
devManager.gemini.apiKey	""	Gemini API key (required)
devManager.gemini.model	"gemini-2.5-flash"	Gemini model used for transcription
devManager.gemini.dictationLanguage	"auto"	Spoken language hint for transcription