Voice Dictation for VS Code on Mac with Hearsy

VS Code is the most popular code editor, but it has no built-in voice dictation. Microsoft's VS Code Speech extension exists, but it requires an internet connection and only works within specific VS Code features. Hearsy takes a different approach. It runs locally on your Mac, transcribes your voice in under 200 milliseconds, and pastes text into any VS Code field -- editor, terminal, search bar, commit input, or source control panel. No extensions needed.

Step-by-step setup

1. 1

   Open VS Code with your project

   Launch VS Code and open the workspace or file you're working on. You can dictate into any text field: the editor, terminal, search bar, commit message input, or extension panels.

2. 2

   Place your cursor in the target field

   Click where you want text to appear. This could be inside a code comment, a markdown file, the terminal prompt, the Git commit message box, or the search bar.

3. 3

   Press your Hearsy hotkey

   Activate Hearsy with your keyboard shortcut (default: Fn twice or a custom hotkey). The menu bar indicator shows that Hearsy is recording.

4. 4

   Speak naturally

   Say what you want to type. Describe a function, explain a bug, dictate a terminal command, or narrate documentation. Hearsy handles punctuation and capitalization for you.

5. 5

   Press the hotkey again to finish

   Tap the hotkey once more. Hearsy transcribes your speech on-device, optionally applies AI enhancement, and pastes the result into VS Code at your cursor position.

Why use Hearsy for VS Code?

Developers spend a surprising amount of time in VS Code writing things that aren't code: comments explaining complex logic, commit messages, documentation, search queries, terminal commands. These tasks are all text input, and typing them out slows you down. Speaking is 3-4x faster.

Hearsy works at the macOS level, not as a VS Code extension. This means it doesn't depend on VS Code's extension API, doesn't need updates when VS Code changes, and works identically in every text field -- including the integrated terminal, where most extensions can't operate. It also means zero configuration inside VS Code itself.

All transcription happens on your Mac. Your voice, your code context, and your project details stay on your computer. For teams with strict data policies or developers working on proprietary software, this is a meaningful difference from cloud-based dictation services. The optional AI enhancement cleans up your speech into polished text locally with Qwen, or through Claude/OpenAI APIs if you prefer.

When to use voice dictation in VS Code

Tips for better results

• ✓

  Dictate markdown documentation directly in .md files. Speaking produces longer, more detailed explanations than most developers write when typing.

• ✓

  Use Hearsy in the integrated terminal for long or complex commands. Instead of typing a multi-flag CLI command, speak it naturally and let Hearsy paste it at the prompt.

• ✓

  For commit messages, open the Source Control panel, click the message input, and dictate a summary of your changes. AI enhancement formats it into a clean, conventional commit message.

• ✓

  Dictate search queries in VS Code's search bar (Cmd+Shift+F). Speaking a natural description of what you're looking for can be faster than constructing a regex pattern from scratch.

• ✓

  Switch to the Whisper engine when writing documentation that includes technical terms from non-English languages. It handles 99+ languages and specialized vocabulary well.

• ✓

  Pair Hearsy with VS Code's Copilot or other AI extensions. Dictate a detailed comment above a function, and the AI assistant uses it as context to generate better code suggestions.