Short answer: Cursor has no built-in voice input, so the best option is a system-wide Mac dictation app like Voice Keyboard Pro. You hold a hotkey, speak your prompt or code comment, and release. The text appears instantly wherever your cursor sits, including Cursor's chat, Composer, and editor panes.
Cursor changed how a lot of developers write software. Instead of typing every line, you describe what you want in plain language and let the AI editor draft, refactor, and explain code alongside you. But there is a quiet bottleneck hiding in that workflow: you are still typing the descriptions. The richer your instructions get, the more you type, and prompts for a non-trivial change can run to several sentences before you even hit enter.
That is where voice input comes in. You think in sentences and you speak roughly three times faster than you type. If the whole premise of Cursor is "tell the computer what you want," it makes sense to actually tell it, out loud, rather than peck it out one key at a time. This guide covers why voice input fits Cursor so well, what to look for in a tool, and how to set it up so dictation feels native to the editor.
Does Cursor have built-in voice input?
No. As of 2026, Cursor does not ship a native microphone button or built-in dictation for its chat, Composer, or inline edit features. You type into those boxes the same way you type anywhere else. There is no first-party "speak your prompt" control inside the editor.
That sounds like a gap, but it is actually good news. Because Cursor is a standard desktop application built on the same foundation as VS Code, any system-wide dictation tool can type into it. You do not need an extension, a plugin, or special Cursor support. You need a voice tool that works at the operating-system level and inserts text wherever your cursor is. Get that right and every text field in Cursor becomes voice-ready at once: the chat panel, the Composer prompt, a code comment, a commit message, a file rename, a search box.
Why voice input makes Cursor faster
The case for dictating to an AI editor is stronger than the case for dictating ordinary documents, and it comes down to how you interact with the tool.
Prompts are conversational by nature. A good Cursor prompt reads like an instruction to a teammate: "Refactor this function so it takes an options object instead of four positional arguments, keep the existing return type, and update the call sites in the same file." That is a sentence you would happily say out loud. Typing it costs you fifteen to twenty seconds; speaking it costs four or five.
The better your prompts, the longer they get. Experienced Cursor users learn that vague prompts produce vague code. The fix is more context, more constraints, more specificity, which means more words. Voice removes the penalty for being thorough. You can ramble out every edge case you care about without dreading the typing.
Flow state survives. Switching between thinking about architecture and hammering out a paragraph of instructions breaks concentration. Speaking keeps your hands near the keys for the precision work (selecting code, accepting diffs, navigating files) while offloading the bulk text entry to your voice. We wrote more about that rhythm in why developers are switching to voice.
The average adult types around 40 words per minute and speaks at 130 to 150. For a tool whose entire interface is "describe what you want," that gap is the whole game.
What to look for in a voice tool for Cursor
Not every dictation tool fits a coding environment. Here is what actually matters when the target is Cursor.
1. It must type system-wide, not in one app
The tool has to insert text into whatever window has focus, including Cursor. Solutions that only work inside a browser or a single notes app are useless here. You want to hold a key, speak, release, and have the words land at the cursor no matter where it is. That same capability lets you dictate in your terminal, your browser, Slack, and your notes without changing tools, as covered in how to dictate in any Mac app.
2. It needs a custom vocabulary for code terms
General dictation engines stumble on the words developers use constantly: library names, framework jargon, your own variable and function names, internal acronyms. A tool that lets you teach it those terms (and fix the ones it keeps getting wrong) is the difference between dictation you trust and dictation you have to babysit. Voice Keyboard Pro's Smart Vocabulary is a personal dictionary with replacement rules, so "next js" reliably becomes "Next.js" and "use effect" becomes "useEffect" every time.
3. Accuracy and punctuation should be hands-off
You should not have to say "comma" and "period" out loud for normal prose. A modern engine adds punctuation, capitalization, and paragraph breaks based on how you actually speak. That keeps your prompts readable without turning dictation into a stilted command language.
4. Low latency
If text takes three seconds to appear after you stop talking, the tool breaks your rhythm and you go back to typing. You want the words on screen in about a second, fast enough that it feels like the editor is keeping up with you rather than lagging behind.
5. Privacy that holds up to a code review
You are going to speak about proprietary code. The tool should not be quietly storing your transcripts. Voice Keyboard Pro's servers store only operational pings, no audio and no transcript content, so your prompts and descriptions are not sitting in a log somewhere.
Setting up Voice Keyboard Pro for Cursor
Voice Keyboard Pro is a native macOS app that lives in your menu bar and types wherever your cursor is. Because it works at the system level, it treats Cursor exactly like any other text field. Setup takes a couple of minutes.
- Install the menu bar app and grant the two macOS permissions it asks for: Microphone (so it can hear you) and Accessibility (so it can type text into other apps, including Cursor).
- Pick a hotkey you can reach without looking. Many developers use a key they would not otherwise press mid-edit, so it never collides with Cursor's own shortcuts.
- Open Cursor and click into any text field: the chat box, the Composer prompt, or directly into a file.
- Hold the hotkey, speak, release. Your words appear at the cursor a beat after you stop. Hit enter to send the prompt, or keep editing.
That is the entire loop. There is nothing Cursor-specific to configure because the app does not integrate with Cursor at all; it integrates with macOS, and Cursor is just another window.
Five ways to use voice inside Cursor
Dictating prompts to chat and Composer
This is the headline use. Click into the chat or Composer box, hold your hotkey, and describe the change. Because you are not paying a typing tax, you naturally give the model more to work with: the file you mean, the constraints you care about, the style you want. Longer, clearer prompts produce better diffs, and voice makes longer prompts effortless.
Writing code comments and docstrings
Comments are prose, and prose is exactly what voice is best at. Place your cursor on a new line, speak the explanation of what a tricky block does, and let the engine punctuate it. Docstrings, JSDoc blocks, and module-level overviews all go faster when you narrate them instead of typing them.
Commit messages
A good commit message is a sentence or two of plain English. Whether you commit from Cursor's source control panel or the integrated terminal, you can dictate the summary and body in the time it takes to think them. Because the app types system-wide, the same trick works at the command line, in your browser, and anywhere else a cursor blinks.
Talking through a refactor
Sometimes the fastest way to plan a change is to think out loud. Open a scratch note or the chat and narrate the approach: what you are trying to achieve, which files are involved, what could break. You end up with a written plan you can paste into Composer or keep as a reference, captured at speaking speed.
Searching, renaming, and navigating
Smaller text inputs benefit too. Dictate a symbol you are searching for, a new file name, or a TODO you want to jot before you forget it. None of these are big enough to feel slow on their own, but they add up over a day, and never reaching for a different input method keeps you in flow.
Voice Keyboard Pro vs. built-in Mac dictation in Cursor
You might wonder why not just use Apple's built-in dictation, since it is free and already on your Mac. It will type into Cursor, but it was not designed for this kind of work, and the friction shows up quickly. Built-in dictation historically capped sessions and could cut off mid-thought, it has no concept of your personal code vocabulary, and it cannot learn the terms it keeps mishearing. For a fuller breakdown of where the stock tool falls short and what a purpose-built alternative does differently, see our guide to a better Apple Dictation alternative and our look at tools that are better than Apple Dictation.
The short version: a dedicated tool gives you a custom vocabulary that learns your codebase, no session limits, consistent punctuation, and a privacy posture that keeps your prompts off the server. Those are precisely the things that matter when the text you are dictating is full of library names and proprietary logic.
Does this work in VS Code, Windsurf, and Zed too?
Yes. Because Voice Keyboard Pro types at the system level, the exact same setup works in VS Code, Windsurf, Zed, JetBrains IDEs, and any other editor on your Mac. There is nothing tying it to Cursor specifically. If you switch editors next month, your voice workflow comes with you, along with the custom vocabulary you have built up. Developers who lean on dictation across their whole stack often pair it with notes apps too, the way we describe in dictation for coding and five ways developers use voice to text daily.
The speed math, concretely
Suppose a focused Cursor session involves twenty prompts, each averaging thirty words. That is six hundred words of instruction. At 40 words per minute typed, that is fifteen minutes of typing spread across the session. At 130 words per minute spoken, it is under five. You just bought back ten minutes per session, and that is before counting the comments, commit messages, and notes you also dictated. Over a working week, the difference is measured in hours, not minutes.
Speed is not even the best part. The bigger win is that you stop rationing your own context. When typing is the cost, you write terse prompts and the model guesses at the rest. When speaking is the cost, you give it everything, and the output gets noticeably better.
Frequently asked questions
Is there a Cursor extension for voice input?
You do not need one. Cursor inherits the VS Code extension ecosystem, but a system-wide dictation app sidesteps extensions entirely by typing into the focused window. That is more reliable than an in-editor plugin because it cannot break when the editor updates.
Will it type code symbols correctly?
For prose like prompts, comments, and commit messages, yes, and that is where voice shines. For dense symbolic code (brackets, operators, exact identifiers) the keyboard is still the right tool. The winning workflow is hybrid: speak the descriptions and explanations, type the precise syntax. Add the identifiers you say often to your Smart Vocabulary so they come out right when you do dictate them.
Does it work offline?
Voice Keyboard Pro uses an advanced AI transcription engine for its accuracy, which needs a connection. The upside is recognition quality that handles accents, technical terms, and background noise far better than older on-device dictation.
How much does it cost?
There is a free tier with daily limits, which is enough to feel the difference in a real Cursor session. Pro is $4.99 a month or $34.99 a year and removes the limits, with the full Smart Vocabulary for your code terms.
The bottom line
Cursor's whole pitch is that you describe what you want and the editor writes it. The fastest way to describe anything is to say it. Since Cursor has no built-in voice input, the best tool for the job is a system-wide dictation app that types into the editor like any other window, learns your code vocabulary, and keeps your prompts private.
Voice Keyboard Pro does exactly that. Install the menu bar app, pick a hotkey, click into Cursor's chat, and start talking to your AI editor the way it was always meant to be used: out loud. Try it free and dictate your next prompt instead of typing it.