The 5 best dictation apps for Mac (2026)
Five speech-to-text apps for the Mac, what each does best, and which ones keep your voice on your own computer.
The best dictation apps for Mac in 2026 are superwhisper (best for AI-polished dictation across devices), MacWhisper (best for transcribing files and meetings), VoiceInk (best open-source dictation with a one-time price), Utter (best for simple, fully on-device dictation with no AI service) and Aiko (best for turning recordings into text offline). If you only dictate short messages in English on an Apple silicon Mac, macOS’s built-in Dictation is often enough.
Disclosure: We make Utter, one of the apps on this list. We’ve tried to place it where it genuinely fits and to describe every app fairly, using each developer’s own site. Prices and features checked 28 October 2026.
Quick comparison
| App | Best for | Price model | Runs locally? | Requires |
|---|---|---|---|---|
| superwhisper | AI-polished dictation | Free plan; Pro monthly, yearly or lifetime | Partly (offline and cloud models) | macOS 14+ |
| MacWhisper | Files, meetings, dictation | Free; Pro one-time | Yes (optional cloud AI) | Not listed |
| VoiceInk | Open-source local dictation | One-time | Yes (optional cloud text enhancement) | macOS 14.4+, Apple silicon |
| Utter | Simple on-device dictation | One-time purchase | Yes | macOS 14+ |
| Aiko | Transcribing recordings | One-time (App Store) | Yes | Recent macOS |
Prices and features checked 28 October 2026 on each developer’s website. Check the official site before you buy.
1. superwhisper: best for AI-polished dictation across devices
superwhisper is one of the best-known AI dictation apps on the Mac. You press a shortcut, talk, and it pastes text into whatever app you’re using. What sets it apart is “modes”: the same speech can come out as a casual chat message, a formal email or a tidy note, because a language model reformats the transcript before it’s pasted.
Standout features
- Predefined modes (Voice, Message, Email, Chat, Formal, Casual and more) plus custom modes with your own instructions.
- “Super Mode” uses what’s on your screen as context, and you can choose between several large language models.
- Offline voice models for transcription, vocabulary for custom terms, and 100+ languages with translation to English on Pro.
- Pro also transcribes uploaded audio and video files, and the Mac app has a meeting assistant.
Limits
- Not everything stays on your Mac: its docs say the default Message mode uses cloud models, so check each mode’s model if privacy matters.
- The developer says offline models run really well only on Apple silicon; Intel Macs work best with cloud models.
- The richest features sit behind Pro, which is a recurring plan unless you buy the lifetime license.
Who it’s for: writers, founders and support staff who want rough speech turned into finished, formatted text, and who use more than one platform.
2. MacWhisper: best for transcribing files and meetings
MacWhisper comes from Jordi Bruin and the small Good Snooze team in the Netherlands. It started as a way to run OpenAI’s Whisper models on a Mac and has grown into a full transcription studio: drag in a file or paste a YouTube link, get a transcript you can edit, translate and export.
Standout features
- Transcribes audio and video files, lectures and podcasts with local models, so sensitive recordings don’t have to leave your Mac.
- Records Zoom, Teams, Webex and other calls without a bot joining the meeting (Pro).
- System-wide dictation with optional AI prompts (Pro).
- Speaker recognition, subtitles (.srt, .vtt), translation and export to .txt, .md, .pdf, .docx and .html.
- Pro adds watched folders, integrations with tools like Notion and Obsidian, and a command-line tool.
Limits
- Dictation and meeting recording need Pro; the free version covers file transcription, microphone recording and basic export.
- AI clean-up through cloud providers (OpenAI, Anthropic, Google and others) means sending text to them with your own API key; you can use local models through Ollama or LM Studio instead.
- It’s a big, feature-rich app. If all you want is to talk into a text field, it may be more than you need.
Who it’s for: journalists, researchers, podcasters and anyone who has hours of recordings to turn into text, and wants dictation in the same app.
3. VoiceInk: best open-source dictation with a one-time price
VoiceInk is a dictation app whose code is public on GitHub under the GPL-3.0 license. The prebuilt app is sold as a one-time license on its website, and it can also be installed with Homebrew. Its pitch is simple: local transcription models running on the Neural Engine, so your voice stays on your Mac.
Standout features
- Local transcription models; the developer says your voice data stays on your Mac.
- Custom dictionary and “Smart Replace” for names, product terms and text expansions.
- Removes filler words and repetition, and can rewrite selected text by voice.
- Switches settings automatically depending on the app or website you’re in, and can use screen context.
- Timestamped quick notes saved to a Markdown journal; an iOS version is available.
Limits
- Apple silicon only, with 8GB of RAM recommended for local models.
- AI enhancement is optional and cloud-based when turned on; the site says only the transcribed text is sent, not your audio.
- The GitHub project isn’t accepting pull requests, so “open source” here means readable and forkable rather than community-developed.
Who it’s for: developers and privacy-minded users on Apple silicon who want local dictation with AI extras and like being able to inspect the code.
4. Utter: best for simple, fully on-device dictation with no AI service
Utter is our app. It’s a small menu bar dictation tool built on Apple’s own speech recognition, with one hard rule: it asks macOS for on-device recognition only and refuses to run a language that would need a server. There’s no language model in the loop and no account. The only network traffic is the app checking our site for updates and activating your license.
Standout features
- Hold to talk. Hold Right Option (or Right Command, Right Control or fn), speak, let go. Prefer hands-free? Use a start/stop shortcut such as Control-Option-Space. Esc cancels.
- Three modes. Clean removes “ums” and “uhs” and fixes capitals; Code keeps casing and builds identifiers (“camel case retry count” becomes
retryCount); Raw gives you exactly what was heard. - Spoken punctuation. Say “new line”, “new paragraph”, “comma” or “period”.
- Your dictionary. Add names and jargon as vocabulary hints, and replacements such as “next js” → Next.js.
- Clipboard-safe pasting. Text is pasted for you and your previous clipboard is put back a moment later.
- Searchable history stored as a local file, plus a count of words dictated and time saved.
What it doesn’t do
- No AI rewriting, tone changes or summaries. What you say is cleaned up by rules, not by a language model.
- No audio or video file transcription, and no meeting recording.
- Only languages your Mac can recognize on-device are offered; macOS may need to download a language’s speech model first.
- Accuracy is Apple’s on-device model. It’s good for everyday speech, but names and technical terms need dictionary entries.
- Mac only. Hearing the hold key in other apps and auto-pasting need Accessibility permission; without it you use the toggle shortcut and press ⌘V yourself.
Who it’s for: people who want a quiet, private upgrade to built-in Dictation for everyday writing and coding, paid for once, with nothing sent to an AI provider.
5. Aiko: best for turning recordings into text offline
Aiko, from prolific Mac developer Sindre Sorhus, is not live dictation. It’s here because many people searching for “offline speech to text on Mac” really want to transcribe a recording, and Aiko does that with very little fuss. Drop in an audio or video file and it runs OpenAI’s Whisper model on your Mac (the large v2 model on macOS).
Standout features
- On-device transcription; the developer states the app collects no data.
- Around 100 languages, and it can translate speech to English.
- Shortcuts support, including batch transcription and Finder Quick Actions.
- Exports text, SRT subtitles, JSON and CSV.
- Options such as “Skip silent parts” and “Reduce repetitions” to work around common Whisper quirks.
Limits
- No live transcription yet, so it can’t type into other apps as you speak.
- Quality varies by language, as with any Whisper-based tool.
Who it’s for: students and anyone with voice memos or lecture recordings who wants a simple, private transcript.
The built-in option: macOS Dictation
Every Mac already has dictation, and for short messages it’s often enough. To turn it on, choose Apple menu › System Settings, click Keyboard in the sidebar, go to Dictation and switch it on. Then pick a shortcut from the Shortcut pop-up menu (for example, pressing the Fn or Globe key twice, or the Microphone key if your keyboard has one), or choose Customize to set your own.
- Start: press your Dictation shortcut, press the Microphone key, or choose Edit › Start Dictation in most apps.
- Stop: press Esc, the Microphone key or the shortcut again. Dictation also stops after 30 seconds without speech.
- Punctuation and formatting: in supported languages, commas, periods and question marks are added automatically. You can also say “new line”, “new paragraph” or a punctuation mark’s name.
Is it on-device? Apple’s Mac Help for macOS 14 Sonoma says that on Apple silicon, dictation requests are processed on your device for supported languages. On Intel-based Macs, or for languages without on-device support, your dictated speech is sent to Apple. Newer versions of macOS let you check in Keyboard settings whether general text Dictation is processed on your device. Dictation into a search box can still send the text to the search provider.
Apple’s current Mac Help also says that with Apple Intelligence on macOS 27, Dictation uses an on-device model to improve spelling, punctuation and capitalization, in English on Macs with M3 or later and at least 12GB of memory.
When built-in is enough: quick replies and notes in English on an Apple silicon Mac. A dedicated app earns its place when you want hold-to-talk anywhere, a custom vocabulary, clean-up of filler words, file transcription, or a guarantee that a language never falls back to a server.
How we picked
- Mac-native and current. Every app is actively sold or distributed for macOS today, from an official site or the App Store.
- Where your voice goes. We only describe on-device or cloud processing as each developer’s own site or docs state it.
- Clear pricing model. Free, one-time or subscription, taken from official pages. We quote exact prices only where the official page showed them clearly.
- Different jobs. Live dictation into any app versus transcribing recordings. We included both because people search for both.
- Verifiable facts only. We left out Wispr Flow, a popular cloud dictation app, because we couldn’t load its official site to confirm current pricing and processing details on the day we checked.
We make Utter. We placed it fourth, as the pick for simple, strictly on-device dictation without an AI service, not as the best app for everyone.
FAQ
What is the best dictation app for Mac?
It depends on what you need. superwhisper is the best fit if you want AI to turn speech into finished emails and messages. MacWhisper is best if you also transcribe recordings and meetings. VoiceInk suits privacy-minded Apple silicon users who want open-source code. Utter is best for simple hold-to-talk dictation that stays entirely on your Mac with Apple’s on-device speech model.
Is there offline speech to text for Mac?
Yes. macOS Dictation runs on-device on Apple silicon Macs for supported languages. Utter uses only Apple’s on-device speech recognition, and VoiceInk, MacWhisper and Aiko run local transcription models. superwhisper offers offline models too, but some of its modes use cloud models by default.
Is Mac dictation free?
Yes. Dictation is built into macOS. Turn it on in System Settings › Keyboard › Dictation and choose a shortcut. Third-party apps add features such as hold-to-talk, custom vocabulary, filler-word removal and file transcription.
Does Mac dictation send my voice to Apple?
On Apple silicon Macs, Apple says dictation is processed on your device for supported languages. On Intel-based Macs, or for languages without on-device support, dictated speech is sent to Apple. Dictating into a search field can send the text to the search provider.
Can I transcribe an audio file on a Mac for free?
MacWhisper’s free version transcribes audio and video files on your Mac. Aiko does the same as a one-time purchase. Built-in macOS Dictation only works live from the microphone, and Utter does not transcribe files.
Which dictation app is best for programmers on Mac?
Look for custom vocabulary and a way to keep code formatting. Utter has a Code mode that keeps casing and builds identifiers such as camelCase and snake_case, plus a dictionary for terms like Next.js. VoiceInk and superwhisper offer custom dictionaries and app-specific modes.
Key takeaways
- For AI-rewritten dictation, superwhisper is the most complete, but check which modes use cloud models.
- For recordings and meetings, MacWhisper (live and files) and Aiko (files only) run transcription on your Mac.
- VoiceInk is the open-source pick, sold as a one-time license, for Apple silicon Macs.
- Utter is the one-time-purchase pick for plain, private dictation: Apple’s on-device model only, no AI service, no account.
- Built-in Dictation is free and on-device on Apple silicon for supported languages; turn it on in System Settings › Keyboard › Dictation.