Can I dictate actual code?

Realistically, no — and we would rather say so than sell you on it. Code is dense with punctuation and casing that speech does not carry: braces, dots, underscores, camelCase, operators. You would spend longer correcting a dictated line than typing it. Dictate the prose around the code and type the code.

Does Cursor or VS Code have built-in dictation?

Neither ships dictation in the editor itself. There are third-party extensions, but nothing native, and an extension only helps inside that editor. VoiceSnap Pro works at the operating-system level, so the same key fills the chat panel, a GitHub textarea in your browser and a Slack message.

So what is it actually good for in an editor?

Prompts for Cursor's chat and Composer, comments that explain why a piece of code exists, commit-message bodies, pull-request descriptions, READMEs and other markdown docs, replies to review comments, and issue descriptions. Broadly: anything that is a sentence rather than a statement.

Does it work in JetBrains IDEs, Zed, Xcode or Neovim?

Yes. VoiceSnap Pro types into the focused text field, so any editor works, including a terminal-based one. In a modal editor make sure you are in insert mode first, or the text will be read as commands.

Will it clash with my keybindings?

Rarely. The dictation key is a right-hand modifier you hold on its own — Right Option, Right Command, Right Control or Right Shift on a Mac; Right Alt, Right Ctrl or Right Shift on Windows — and editors seldom bind one by itself. If yours does, pick another in Settings → Shortcuts.

Does dictating a longer prompt actually improve what Cursor writes?

In practice, yes — for the same reason it improves any AI output. The model can only work with the constraints you gave it: the file, the existing pattern, the edge case, the thing not to touch. Typing all that is expensive, so people do not; saying it takes forty seconds.

Is my code sent anywhere?

No. VoiceSnap Pro has no access to your files, your repository or your editor's contents. It transcribes what you say while the key is held and inserts the text at the cursor, and nothing you dictate is used to train AI models.

Code editors · Dictation guide

How to dictate in Cursor

Hold one keyboard shortcut, speak, and punctuated text lands in Cursor — and in every other app on your Mac or PC.

Download free Try it free in your browser

Cursor is in Cursor. You hold the shortcut and say:

“hey sam thanks for the detailed write up i think we should ship the fix on friday and hold the rest for the next release”

What appears in Cursor

Hey Sam, thanks for the detailed write-up. I think we should ship the fix on Friday and hold the rest for the next release.

Punctuation and capitals added, text pasted where the cursor already was.

Let's be straight about this one: dictation is bad at code. Code is punctuation-dense — braces, dots, underscores, camelCase — and saying it out loud is slower and more error-prone than typing it. What dictation is genuinely good at in an editor is everything around the code: the prompt in Cursor's chat, the comment explaining why a function exists, the commit message body, the pull-request description, the README paragraph, the reply to a review comment. That is a large share of what a developer types in a day, and almost none of it is syntax.

Punctuation you never say

Inferred from how you speak, so you never say “comma”.

Your last dictation, again

One shortcut pastes it again, wherever you are.

Custom vocabulary

Names, jargon and acronyms, taught once.

Saved and searchable

Keep a text copy in optional local dictation history, or use the voice-note shortcut to save an idea across devices.

What people actually dictate in Cursor

The things worth using it for on day one, rather than a list of everything it can technically do.

Prompts for Cursor's chat and inline edit

The quality of what an AI editor gives back tracks how much of the problem you described. Say the constraint, the file it has to work with, the pattern the codebase already uses and what you do not want it to touch. Forty seconds of talking, instead of the two-line prompt you would have typed.

Comments that explain why

Nobody skips the “why” comment because they do not value it; they skip it because typing a paragraph mid-flow breaks concentration. “We retry twice here because the upstream API returns 502 on a cold start” is a sentence, and sentences dictate perfectly.

Commit messages and PR descriptions

The commit body and the PR description are prose, and they are the parts most likely to be one line because writing them is a chore. Dictate the reasoning, the trade-off and the thing a reviewer should look at first.

READMEs, docs and review replies

Markdown prose, migration notes, an answer to a reviewer who asked why you did it that way. All ordinary writing that happens to live inside an editor or a GitHub textarea.

How to dictate in Cursor, step by step

Five minutes of setup, then the same held key forever.

Step 1
Install VoiceSnap Pro and pick a shortcut your editor does not use
Menu bar on macOS, tray on Windows, and a sign-in with Google or an emailed code. Editors bind an enormous number of key combinations, which is why the dictation key is a right-hand modifier you hold on its own: editors rarely bind one by itself.
Step 2
Choose the field, not the file
Cursor's chat panel, the Composer input, a comment line in the editor pane, a markdown file, the commit-message box in the source-control sidebar, or the integrated terminal. Click deliberately: the cursor is what decides where the text goes.
Step 3
Hold the shortcut and speak in prose
Sentences, not syntax. Punctuation arrives automatically, so a dictated comment is already a readable sentence; press Return yourself between the paragraphs of a PR description.
Step 4
Release and fix the casing
Identifiers are the one thing to expect to touch up. A spoken “handle submit async” arrives as three words, not as handleSubmitAsync. Fix it in a second, or type the identifier and dictate the sentence around it.
Step 5
Teach it your stack
Repo names, library names, service names and internal acronyms go into your custom vocabulary. It is for words, not for syntax — it will not turn speech into camelCase, but it will stop “Postgres” and your service names coming back wrong.

Tips for dictating in Cursor

The things people work out in the second week, collected here for the first day.

Type the commit subject line, dictate the body. Subjects are short and conventional; bodies are where the explanation lives.
Vim mode: make sure you are in insert mode before you dictate, or the text arrives as a string of commands.
The integrated terminal is a text field too — useful for a long commit message, much less useful for a command full of flags.
For an identifier inside a dictated sentence, say the words and fix the casing after, or type the identifier first and dictate around it.
Cursor's chat, the Composer and the editor pane are separate fields. Click the one you mean before you hold the key.
Every prompt you dictate is kept in your dictation history, which turns into a surprisingly useful log of what you asked an AI to do this week.

Cursor dictation FAQ

: Realistically, no — and we would rather say so than sell you on it. Code is dense with punctuation and casing that speech does not carry: braces, dots, underscores, camelCase, operators. You would spend longer correcting a dictated line than typing it. Dictate the prose around the code and type the code.
: Neither ships dictation in the editor itself. There are third-party extensions, but nothing native, and an extension only helps inside that editor. VoiceSnap Pro works at the operating-system level, so the same key fills the chat panel, a GitHub textarea in your browser and a Slack message.
: Prompts for Cursor's chat and Composer, comments that explain why a piece of code exists, commit-message bodies, pull-request descriptions, READMEs and other markdown docs, replies to review comments, and issue descriptions. Broadly: anything that is a sentence rather than a statement.
: Yes. VoiceSnap Pro types into the focused text field, so any editor works, including a terminal-based one. In a modal editor make sure you are in insert mode first, or the text will be read as commands.
: Rarely. The dictation key is a right-hand modifier you hold on its own — Right Option, Right Command, Right Control or Right Shift on a Mac; Right Alt, Right Ctrl or Right Shift on Windows — and editors seldom bind one by itself. If yours does, pick another in Settings → Shortcuts.
: In practice, yes — for the same reason it improves any AI output. The model can only work with the constraints you gave it: the file, the existing pattern, the edge case, the thing not to touch. Typing all that is expensive, so people do not; saying it takes forty seconds.
: No. VoiceSnap Pro has no access to your files, your repository or your editor's contents. It transcribes what you say while the key is held and inserts the text at the cursor, and nothing you dictate is used to train AI models.