Granola Voice Typing Is Here: Free, Unlimited Dictation on Mac with Multilingual Support

Hi, Hoda here!

Granola has added a dictation feature called Voice typing to its Mac desktop app. Announced on October 29, 2026, it’s currently in beta and available on all plans, including the free one, with no word limit.

Granola is best known for meeting notes, but Voice typing is a separate feature. It lets you dictate text into any app on your Mac, including Gmail, Slack, ChatGPT, and Claude.

In this article, I’ll walk through Granola’s voice typing: how to use it, its main features, pricing, privacy, and how it compares to Wispr Flow, including what I found when I tried it myself.

What Is Granola’s Voice Typing?

Voice typing lets you hold a key and speak, and Granola turns what you said into polished text and pastes it wherever your cursor is.

It’s more than plain transcription. The text is cleaned up automatically:

• Removes filler words (“um,” “uh”) and accidental false starts
• Adds punctuation
• Splits text into paragraphs by topic
• Turns spoken lists into bullet points or numbered lists

The basic idea is to make your words easier to read without changing what you said. For example, if you tell ChatGPT, “Write a reply to Sam,” Granola won’t respond to that request. It simply types out what you said.

Note that Voice typing is separate from meeting transcription. It doesn’t start a new meeting, and it works in apps other than Granola.

How to Use It

There are two ways to use Voice typing: “Hold to talk,” where you speak only while holding a key, and “Hands-free,” which keeps going after you let go.

Initial Setup

1. Click your workspace name at the bottom left of Granola and open Settings
2. Open Voice typing and turn on Enable Voice Typing
3. Follow the on-screen prompts to grant Granola Accessibility permission in macOS
4. Allow microphone access and test that it works

Accessibility permission is used to detect your shortcut, read context from the app you’re typing in, and paste text into other apps.

Hold to Talk

1. Click the text field where you want to type
2. Hold down the shortcut (the default is the Fn key) and speak
3. Release the key when you’re done, and the cleaned-up text is pasted in

While you’re speaking, a bar with a waveform appears at the bottom of the screen.

Hands-Free

For longer passages, double-press the shortcut to start hands-free mode. Press the shortcut again, or end it from the bar at the bottom of the screen, to finish.

A single voice typing session can last up to 20 minutes. When it hits 20 minutes, everything up to that point is pasted and the session ends automatically.

To stop partway through, press the Esc key to cancel without pasting anything.

Changing the Shortcut

You can change the shortcut to any key or key combination under Settings > Voice typing > Shortcut. However, common editing shortcuts such as Cmd+C, Cmd+V, Cmd+Z, and Cmd+Tab can’t be assigned.

Main Features

Language Settings (Japanese and Multiple Languages Supported)

Under Settings > Voice typing > Language, you can choose the language you speak. Japanese is among the options, and you can set more than one language at a time, such as Japanese and Spanish.

If you switch languages while speaking, choose Auto-detect. Your words are entered in the language you spoke, and nothing is translated.

Dictionary

This lets you register names, product names, internal terminology, and other words you want spelled a certain way. Add them under Settings > Voice typing > Personalization > Dictionary.

• Each entry can be up to 50 characters and 6 words
• It’s meant to help with recognition, and it can’t replace one word with another
• It’s separate from the jargon list used for meeting transcription, so register terms in both places if you want them in both

Auto-add Words

If Granola gets a name or term wrong and you correct it yourself right after it’s pasted, the correct spelling is automatically added to your dictionary. You’ll see a notification in the bar at the bottom of the screen, and you can undo it.

Only spelling corrections to names and terms are learned. Rephrasing a sentence doesn’t affect your dictionary. If you don’t want this, turn off Auto-add words under Settings > Voice typing > General.

Screen Context (Correct Spelling from What’s on Screen)

With Screen context on, Granola reads the text visible in the app you’re typing into, so it can spell things like an email recipient’s name or the names of people in a Slack thread correctly. It’s on by default.

The shape of your text also adapts to the type of app:

• Email: separates the greeting and sign-off, and groups the content into paragraphs
• Chat apps like Slack: short, conversational text
• Code editors and terminals: preserves the capitalization of technical terms
• Single-line fields such as subject lines and search boxes: no line breaks or final period
• When continuing mid-sentence: blends in with the text before and after the cursor

Styles (Per-App Tone)

Under Settings > Voice typing > Personalization > Styles, you can choose a writing style for each type of app. Styles only change capitalization, punctuation, and spacing. The words you said stay the same.

The default is Automatic, which picks punctuation to match what you said. Web apps in a browser are identified by URL, so Gmail in Chrome is treated as email and Slack as work chat.

Screen context must be on to use Styles. If you turn it off, the Styles settings are grayed out.

History

You can review your past voice typing under Settings > Voice typing > History. The settings screen shows your five most recent entries, and opening “History” shows everything grouped by day.

• Click an entry to copy its contents
• Delete entries one at a time
• Use “Clear all” to delete everything (this can’t be undone)

If an input fails, for example because your connection drops, the audio is kept for up to 24 hours so you can retry without speaking again.

Where Your Data Is Stored (Cloud or Local)

You can choose where your history is stored under Settings > Voice typing > Data and privacy.

• If you turn on Store data in the cloud, your history syncs across devices. If it’s off, your history stays only on that Mac

What stays on your Mac is set under Local data storage, with three options:

With “Never store data locally,” you won’t be able to retry failed inputs.

Pricing: Unlimited Words, Even on the Free Plan

Voice typing is available on all Granola plans, including the free Basic plan, at no extra cost and with no word limit.

That said, it’s currently in beta. We don’t know yet whether pricing and limits will stay the same once it’s officially released.

Supported Platforms

• Mac: available in beta (in the Granola desktop app)
• Windows and mobile: listed as “coming soon,” with no timeline announced

You’ll need to grant macOS Accessibility permission and microphone access. It uses your Mac’s default microphone, and you can switch to a different one. Granola says built-in and wired microphones are the most stable, and Bluetooth microphones tend to be less reliable.

Mute other audio is also on by default. It automatically mutes your Mac’s sound while you speak so music or video doesn’t get mixed in.

Privacy and Data Handling

When you use voice typing, Granola sends the following data to its transcription and AI providers:

• Your audio
• Your dictionary contents and selected language
• If Screen context is on, part of the text in the app you’re typing into

In other words, your audio is not processed entirely on your Mac. Granola says its providers don’t use voice typing content to train AI models.

Screen context reads text through macOS accessibility features and doesn’t take screenshots. It doesn’t read password fields, and the text it reads is sent only for that input and isn’t saved in your history. If you’re concerned, you can turn off Screen context (though Styles and per-app formatting won’t be available).

Granola only listens from the moment you press the shortcut until you finish or cancel your input.

How It Compares to Wispr Flow

With Voice typing, Granola now competes with Wispr Flow in similar territory. Wispr Flow grew up as a dictation tool, and in August 2026 it also launched meeting notes for Mac. Each company is moving into the other’s territory.

Wispr Flow’s free tier limits have been described differently more than once recently, so check the official pricing page for the latest cap.

Background: Meeting Notes and Dictation Are Converging

According to TechCrunch, Granola’s announcement is part of a broader trend in voice AI, where apps are becoming more alike. Meeting-notes apps and dictation apps used to be separate categories, but they’re now moving toward offering both.

• Granola: from meeting notes to dictation
• Wispr Flow: from dictation to meeting notes
• Fireflies, Speechify, Monologue, and others are also becoming apps that handle both meeting notes and dictation

TechCrunch’s article puts Granola’s valuation at $1.5 billion and Wispr Flow’s at $2 billion. What each company does with spoken content next looks like the next battleground.

Pros and Cons

Summary

Here are the key points about Granola’s voice typing:

• Available in beta in the Mac desktop app. It’s separate from meeting notes and works in any app
• Two modes: “Hold to talk” (speak while holding a key such as Fn) and “Hands-free” (double-press to keep going)
• Rich features, including multiple languages (Japanese included), a dictionary, Screen context, per-app styles, history, and storage settings
• Free and unlimited on all plans (as of the current beta)