Voice Typing for Cursor: A Practical Workflow Guide

TL;DR: Voice typing for Cursor is a perfect fit because almost everything you type into Cursor is natural language, and typing it is the bottleneck. With Infina you can speak thousands of words of instructions a day, keep Cursor's agent (and a terminal, and Claude Code) busy in parallel, and do more in less time. Infina runs the whole loop by voice: dictate, send, and switch apps, with on-device transcription by default. It's $8.25 a month, billed annually at $99, with unlimited words and voice commands, and a 7-day refund.

The workflow below assumes the tool can finish the loop. This is what that means. Infina runs on three spoken triggers and there is no fourth. Say "type" and a sentence and it lands as typed text in whatever app is in front of you. Say "send it", or just "enter", and it presses Enter so the prompt actually goes. Say "open Cursor" or "open Slack" and your computer switches apps for you.

Nothing is held down and nothing is pressed. There is no push-to-talk key to find and no Enter key to reach for two hundred times a day, which is the part every other tool in this category leaves to your hands. Wispr Flow, Superwhisper and the rest all stop at inserting text, which leaves the two keypresses around every prompt exactly where they were.

Why voice typing for Cursor works better than in a normal editor

A classic editor is keystroke-dense: symbols, operators, precise syntax. Miserable to dictate. Cursor inverted that.

Per Cursor's own docs, its core surfaces take plain English:

• Chat: questions about the codebase, debugging conversations.
• Composer / Agent: multi-file instructions, like "add auth to these routes and update the tests."
• Inline edits: select code, describe the change you want.

That means the text you produce all day in Cursor is prose, not code. And prose is exactly what dictation is good at: you speak your intent, the model writes the syntax. Speaking is also about three times faster than typing, the core argument in dictate prompts to AI.

The more you lean on Cursor's AI, the smaller the share of your day that deserves a keyboard.

There's a second effect: spoken instructions tend to be longer and richer than typed ones, because talking is cheap. "Refactor this" becomes "refactor this to use the shared client, keep the public API the same, and don't touch the tests." Richer instructions get you better agent runs.

Dictating into chat and Composer

The mechanics are simple, which is the point. With Infina:

1. Click into Cursor's chat panel or Composer input.
2. Hold Option (⌥), speak, release. The text lands at the cursor.
3. Press Enter to send.

Because Infina types at the OS level into whatever's focused, there's nothing Cursor-specific to install. Chat, Composer, inline-edit boxes, the built-in terminal, commit message fields: same gesture everywhere. The same is true if your editor of choice is VS Code or Windsurf; Infina doesn't care which app is focused.

Transcription runs on your Mac by default (Apple Silicon), works offline, and your audio never leaves the device.

Two habits worth building for Cursor specifically:

• Reference code by selection, not by voice. Don't dictate identifiers character by character. Select the code or @-mention the file with a click, then describe the change verbally.
• Dictate the whole instruction in one breath. Composer does better with one complete brief than three fragmented follow-ups. Speaking makes complete briefs cheap.

Iterating on agent output by voice

The real time sink in Cursor isn't the first prompt. It's the loop after it: the agent produces a diff, you read it, and you respond.

Typed, each response is a small chore, and the friction quietly pushes you toward accepting mediocre output. Spoken, the loop tightens. Eyes on the diff, hold Option, and say what you see:

• "The middleware change is right but revert what you did to the config file."
• "Good, now do the same pattern for the other two endpoints."
• "You misunderstood: the cache should be per-user, not global. Try again."

None of this needs punctuation or polish; Cursor's models parse conversational corrections fine. You review more, settle less, and ship faster. This style of working, describing changes to the agent by speaking, is vibe coding by voice.

The hands-free loop for long agent sessions

Push-to-talk still ties every instruction to the keyboard: hold the key, then press Enter, then Cmd-Tab when the agent starts a long run. During heavy sessions (agent grinding in Cursor, a build in the terminal, maybe a Claude Code session on the side) Infina's hands-free mode removes those last touches:

1. Double-tap Cmd (⌘) to enable hands-free. Listening runs on-device; nothing is recorded or sent while it waits.
2. Speak a sentence that starts with "type", then your instruction to Cursor. Infina types it in.
3. Say "send it": Enter is pressed for you.
4. While the agent runs, say "open Terminal", check the build, queue a command there, and switch back. All by voice, from a few feet away.

This is how one person keeps multiple agents busy at once. It's the mode for reviewing on a second screen, standing, or pacing while the agent works.

One honest note: hands-free is our newest surface and labeled experimental, so it ships off by default and it's happiest in a reasonably quiet room. Push-to-talk always works as the fallback. The full concept is laid out in hands-free voice prompting.

When raw dictation is enough, and when it isn't

Base Infina deliberately outputs raw text: on-device transcription with fast rule-based cleanup, no LLM rewrite. For 90% of what you type into Cursor, that's the correct trade:

• Raw is fine: chat, Composer instructions, inline-edit requests, terminal-ish notes. Cursor's AI doesn't care about your commas, and raw means no cloud round-trip and no latency tax.
• Polish matters: text that humans read verbatim, like code comments, commit messages, PR descriptions, README prose. Dictating those raw means cleaning them up by hand.

If you dictate a lot of human-facing text, the clean path is Infina's optional $5/month cloud add-on (billed annually at $60): sharper cloud transcription plus LLM cleanup and more languages, on top of the yearly plan ($8.25 a month, billed annually at $99) you already have. That is the exact job subscription tools charge $15/month forever for.

Honest alternatives for Cursor

• macOS built-in dictation: free, and it types into Cursor like any text field. Triggered from the keyboard, mixed accuracy on technical vocabulary, but the zero-cost way to test whether the habit sticks.
• Wispr Flow: has a Cursor extension and covers Mac, Windows, and phones. The trade-offs: it's cloud-only (no offline mode) and priced at $15/month, which is $180 a year, versus Infina's on-device default, a flat $99/year with unlimited words and commands, and a $5/month cloud add-on for LLM cleanup when you want it. Full comparison: Wispr Flow vs Infina.
• Infina: raw, private, on-device dictation plus the piece nothing else in this list has: voice sending and voice app-switching for hands-free agent sessions. $8.25 a month, billed annually at $99 (at the time of writing), 7-day no-questions refund, optional $5/month cloud add-on; details on pricing. For how the wider field prices dictation, see how dictation app pricing compares.

For occasional dictation, free is fine. For hours of daily agent work, the hands-free loop is the difference, and $8.25 a month, billed annually at $99, covers unlimited use of it.

FAQ

Can I dictate into Cursor's chat and Composer? Yes. System-level dictation tools type into any focused text field, so chat, Composer, inline-edit boxes, and Cursor's terminal all work with the same hold-Option gesture. No extension required.

Do I need an extension to talk to Cursor? Not with a system-wide tool like Infina or macOS dictation; they type wherever your cursor is. Wispr Flow offers a dedicated Cursor extension; that's their integration model, tied to their cloud subscription.

Does my dictation need punctuation for Cursor's AI? No. Chat and Composer handle conversational, unpunctuated speech well; that's why raw dictation is the right default for prompting. Punctuation matters when you dictate comments, commits, or docs that humans read.

Can I send a Cursor prompt without touching the keyboard? With Infina's hands-free mode, yes: speak a sentence that starts with "type", then say "send it". It types the message and presses Enter, and "open [app]" moves you between windows by voice.

Is my code or audio sent to the cloud when I dictate? Not by default with Infina. Transcription runs entirely on your Mac and nothing is stored, so it works offline too; cloud processing exists only as the optional add-on. See on-device dictation for Mac.

What does Infina cost for Cursor users? Same as for everyone: $8.25 a month, billed annually at $99 (at the time of writing), with unlimited words and voice commands, all updates included, 7-day money-back guarantee. The optional cloud add-on for polished output and more languages is $5/month billed annually.

The bottom line

Cursor already turned coding into describing changes in English; voice typing just moves that English through a faster pipe. Speak your instructions and you produce more of them, in less time, with richer detail.

Start free with macOS dictation to test the habit. Then pick by your real bottleneck: if it's polished prose everywhere, Infina's $5/month cloud add-on handles that on top of the yearly plan ($8.25 a month, billed annually at $99).

If it's long agent sessions where you want to instruct, send, and switch apps without leaving review mode (hands-free, on-device, unlimited), that's precisely the loop Infina was built for. $8.25 a month, billed annually at $99, risk-free for 7 days.

---

Picking the tool that fits this workflow

The workflow above assumes a tool that can finish the loop. Here is how the field lines up against it.

The tools this page covers (Wispr Flow, Apple Dictation) are at the top. The rest of the field follows, because the right answer is sometimes none of the above.

The part that is actually different: zero keypresses

The mainstream dictation apps (Wispr Flow, Superwhisper, MacWhisper, VoiceInk, Apple Dictation) are all push-to-talk: you hold or press a key to start talking, and you press Enter to send. That is two physical acts on every single prompt. A hundred prompts in a day is two hundred keypresses that exist only to start and stop the tool. Infina removes both of them. Nothing is held, nothing is pressed, and the same loop runs on three spoken words.

There are three triggers and there is no fourth. "type" types, "send it" sends, "open" opens and switches apps. The send trigger answers to either wording, "send it" or a plain "enter", whichever you find yourself saying. That is the entire vocabulary: no command language, no grammar, no custom scripts to write, nothing to memorize beyond those three words.

When Infina is not the answer. If you need dictation on an iPhone, an iPad, an Android phone or Linux, Infina is a Mac and Windows desktop app and there is no phone version. Wispr Flow covers phones.

Everywhere else, the difference is the loop. Every other tool in that table types words into a box and stops, leaving your hands to press Enter and pick the next window. Infina finishes it: say "type" and your sentence, say "send it" (or just "enter"), say "open Cursor", and keep going without a key.