TL;DR: Voice typing for VS Code has a genuinely good free option: Microsoft's official VS Code Speech extension dictates into the editor and Copilot Chat with local processing (checked July 4, 2026). Its limit is the app boundary: it works only inside VS Code, one push-to-talk dictation at a time, hands on the keyboard. Infina works in every app on your Mac and runs the whole loop hands-free: lean back two feet from the desk, say "type" plus your prompt, say "send", then "open Terminal" by voice while the agent works. On-device by default, $99 once, 7-day refund.

The built-in answer: VS Code Speech

Let's be straight about this, because plenty of dictation vendors pretend otherwise. VS Code has an official, free voice extension, and it is good at what it does.

VS Code Speech is published by Microsoft. As of July 4, 2026, per its Marketplace listing:

  • It adds speech-to-text to VS Code's chat interfaces, so you can speak to Copilot Chat through a mic icon.
  • It offers dictation into the editor via a Voice: Start Dictation command with a keyboard shortcut.
  • Processing is local: the listing states no internet connection is required and voice audio is processed on your computer.
  • It is free, with support for 26 languages.

If your entire working life happens inside VS Code and you dictate occasionally, install it. It costs nothing and it is the honest baseline every paid tool should be measured against.

Where voice typing for VS Code hits the editor's edge

Here is the structural limit: VS Code Speech is a VS Code extension, so it ends where VS Code ends.

Your day does not. A developer's text output is scattered across the terminal outside the editor, the browser (docs, GitHub reviews, ChatGPT or Claude tabs), Slack, commit tools, Notes, and increasingly a Claude Code or Codex session in a separate terminal window. None of that can hear you.

The second limit is the interaction model. VS Code Speech is push-to-talk bound to the editor: press the shortcut or click the mic, speak, then hit Enter yourself. Every single dictation starts and ends with your hands on the keyboard.

That model breaks down exactly where voice matters most in 2026: long AI sessions. When an agent is grinding through a task, the valuable loop is prompt, send, switch to another window, check something, come back, correct. A dictation feature that is chained to one app and one hotkey cannot run that loop.

Going system-wide: one gesture, every app

Infina types at the OS level into whatever field has focus, so one gesture covers your entire Mac:

  1. Click into any text field: Copilot Chat, the editor, a terminal, Slack, a browser.
  2. Hold Option (⌥), speak, release.
  3. The text lands at your cursor.

Transcription runs fully on your Mac by default (Apple Silicon), works offline, and your audio never leaves the device. Like VS Code Speech, it is local; unlike VS Code Speech, it is local everywhere.

That "everywhere" matters more than it sounds. Speaking your prompts instead of typing them is roughly a three-times speed upgrade, and the win compounds when it applies to every AI surface you use, not one. That is the full argument in dictate prompts to AI.

It also covers the VS Code corners the extension reaches and the ones outside it with the same muscle memory: dictate code comments, commit messages, PR descriptions, and terminal commands without switching tools.

The hands-free loop VS Code Speech cannot run

This is Infina's category difference, and no editor extension has an answer to it.

Turn on hands-free mode with a double-tap of Cmd (⌘). From then on, from two or three feet away:

  1. Say a sentence starting with "type": "type explain this error and suggest a fix." Infina types it into Copilot Chat.
  2. Say "send". Enter is pressed for you.
  3. While Copilot or your agent works, say "open Terminal" to watch the build, or "open Windsurf" to brief a second agent, then "open Visual Studio Code" to come back.

No hotkey per dictation, no reaching for Enter, no Cmd-Tab. You can review a diff on a second monitor, eat lunch, or pace behind the chair while keeping two or three agents fed. The same loop powers voice typing for Cursor and voice typing for Windsurf if you split time across editors.

Honest note: hands-free is our newest surface, labeled experimental, and off by default. It prefers a reasonably quiet room, and hold-Option push-to-talk is always there as the fallback.

Raw by default, polished when you want it

Base Infina outputs raw text: on-device transcription with fast rule-based cleanup, no LLM rewrite, no cloud round-trip. For prompts, chat, and terminal commands, raw is the right trade because the AI does not care about your punctuation.

For prose humans read verbatim, the optional $10/month cloud add-on brings sharper cloud transcription, LLM-polished punctuation and formatting, and more languages. Subscription dictation apps charge $15/month forever for that; with Infina it is an optional layer on an app you own outright.

The license is $99 one-time as of July 4, 2026, includes every 1.x update, and carries a 7-day no-questions refund. Details on pricing.

Which one should you use?

Honestly: possibly both.

  • VS Code Speech: free, official, local. The right tool if you dictate rarely and only inside VS Code.
  • Infina: the right tool when voice becomes a workflow. Every app on your Mac, on-device by default, and the hands-free prompt, send, and switch loop that no editor-bound extension can offer. One $99 purchase instead of a subscription.

They do not conflict. Plenty of people install the free extension, feel the speed difference, and then want that speed in the other 60% of their day.

FAQ

Does VS Code have built-in voice typing? Yes. Microsoft ships the free VS Code Speech extension, which adds dictation in the editor and voice input for Copilot Chat, with local processing (checked July 4, 2026). It works only inside VS Code, so other apps need a system-wide tool.

Is VS Code Speech free and does it work offline? Yes. As of July 4, 2026 the extension is free, and its Marketplace listing states voice audio is processed locally on your computer with no internet connection required.

How is Infina different from the VS Code Speech extension? Scope and hands. Infina types into any app on your Mac (browser, terminal, Slack, other editors), not just VS Code, and its hands-free mode runs the whole loop by voice: say "type" plus your words, say "send" to press Enter, say "open" plus an app name to switch. VS Code Speech is push-to-talk inside one app.

Can I use voice with GitHub Copilot Chat? Yes, two ways. VS Code Speech adds a mic to Copilot Chat directly, and Infina dictates into the same chat box from the OS side with hold-Option, which also works in every other AI tool you use.

Can I dictate code comments and commit messages? Yes. Infina types into the editor, the commit message box, and any Git client the same way. Raw on-device output is fine for prompts; for polished human-facing prose, the $10/month cloud add-on applies LLM cleanup automatically.

Is my audio sent to the cloud when I dictate? Not by default with Infina. Transcription runs entirely on your Mac and works offline; your audio never leaves the device. Cloud processing exists only as the optional paid add-on.

The bottom line

Voice typing for VS Code is a solved problem inside the editor: VS Code Speech is free, official, and local, and we are not going to pretend otherwise.

But the editor's edge is the wrong place for your voice workflow to stop. Infina gives you the same on-device privacy across every app on your Mac, then adds the thing nothing else in this article does: a fully hands-free loop where you speak a prompt, say "send", and switch apps by voice from across the room.

$99 once, no subscription, 7-day money-back guarantee. Install the free extension today if you like, and when voice becomes how you work, Infina is how it covers the rest of your Mac.