How to Dictate Into VS Code on Windows: Built-In Voice vs PipeVoice

VS Code already has voice input, so you do not need a third-party tool to dictate into it. PipeVoice is worth adding when you want the same hotkey and the same engine in VS Code and everywhere else. This page covers what VS Code does natively, then how PipeVoice behaves inside it.

What VS Code's own voice input does

Straight from the VS Code documentation:

• Dictation starts with Ctrl+I in the chat window (press again to stop), Ctrl+Alt+V in the editor, and a command-palette command in the terminal.
• On desktop, the default speech recognition model processes microphone audio on your device, and it supports Windows on x64 and Arm64.
• The chat input also has an experimental Voice Mode for hands-free conversation, which Microsoft says "might change or be removed".
• In chat, you press Enter or select the Send button to submit.

If you only ever dictate inside VS Code, that may be all you need.

Where a system-wide tool helps

VS Code's dictation lives in VS Code. PipeVoice is a separate push-to-talk app: hold Ctrl+\, speak, release, and it types into whatever window has focus. You keep one hotkey for VS Code, your browser, and your commit messages, and you choose the engine (Deepgram, OpenAI Whisper or local Whisper) instead of taking the default.

Matching VS Code with a per-app profile

PipeVoice recognises the focused program by its executable name. Visual Studio Code is code.exe in PipeVoice's app picker, so a profile for code.exe only applies while VS Code is in front. A profile points at a Voice, and a Voice can override the engine, AI cleanup style, auto-Enter, and whether output is typed, pasted or sent to the clipboard.

The built-in Code / Prompt Voice uses the "prompt" cleanup style, which reorders rambling speech into clear sentences without changing its meaning. That suits long Copilot chat requests.

Chat: auto-Enter or review first

Because Enter submits a chat message in VS Code, PipeVoice's auto-Enter setting will send your prompt the moment it finishes typing. Leave it off for a profile if you like to read a long prompt before sending. Either way, ending a dictation with "send it", "press enter", "hit enter", "enter key" or "send message" types the text and then presses Enter once.

Terminal: typing vs pasting

By default PipeVoice types keystrokes. If you switch a profile's output to paste, it copies the text, sends Ctrl+V, then restores your previous clipboard. VS Code documents Ctrl+V as the paste shortcut in its integrated terminal on Windows, so paste mode works there. The terminal toggles with Ctrl+`. For shell-specific behaviour see dictating into the terminal on Windows.

Set it up in three steps

1. Download PipeVoice for Windows, run it, and choose an engine.
2. Open the app profiles window and click into VS Code. It shows the exact program name to match on; add code.exe and pick a Voice.
3. Click into a chat box, hold the hotkey, speak, release.

Using Claude Code or Cursor instead? See dictating into Claude Code.