Superwhisper vs Typeless on Mac in 2026: price, privacy and which dictation app fits

Superwhisper vs Typeless: what each app is built to do

Both apps are built for the same basic job. You hold a key, talk, and clean text lands where your cursor sits. The differences show up in three places: where the speech model runs, what each plan costs, and how much control you get over the output. Read this page with your own work in mind. A lawyer dictating client memos needs different things than a developer dictating commit notes.

Superwhisper leans toward control. Its free plan runs local Whisper models with unlimited dictation. You can add a custom vocabulary and text replacements, and you can pick modes that shape the output. Typeless leans toward polish. It removes filler words and repeats, keeps your final wording after you correct yourself, and adapts its style to each app.

Neither app is the right answer for everyone. Superwhisper asks you to make a few choices up front, such as which model to run. Typeless asks you to trust its cloud and its cleanup. Both trade-offs are reasonable. You just need to know which one you are making.

How much Superwhisper and Typeless cost in 2026

Superwhisper's billing plans page lists a Free plan at $0. Pro is $8.49 per month, $84.99 per year, or $249.99 as a one-time lifetime purchase. The docs say every plan has the same features, and only billing differs. So you are not paying for a better model. You are choosing how you pay.

Typeless lists three tiers on its pricing page. Free is $0 with a limited number of words. The cap is not published, and third-party blogs give conflicting figures, so do not rely on those. Pro is $12 USD per member per month billed yearly, or $30 when billed monthly. Enterprise is custom.

Here is a simple way to compare. Look at the billing period you would actually use. Superwhisper's monthly Pro is $8.49. Typeless's monthly-billed Pro is $30 per member. If you plan to stay for a year, Superwhisper's annual option is $84.99 total, and Typeless's yearly option is $12 per member per month. Those are different units, so compare the total you would pay, not the headline.

Check each pricing page before you buy. Superwhisper's /pricing URL returned a 404 when checked on 2026-11-08. The figures above come from its billing plans page in the docs. Its homepage also shows a label of $849, which looks like a display error for $8.49. Prices change, so confirm them at checkout.

Free plans: what you actually get for $0

Superwhisper's free tier is the more generous one for dictation alone. Unlimited dictation with local Whisper models is included. So are custom vocabulary, text replacements, and up to two modes that do not use AI processing. Meetings only get free trial words, so do not count on free meeting notes.

Typeless's free tier is a taste of the core product. It is capped by a word limit that is not on its pricing page. The page says Limited words and nothing more. If you dictate a lot, you may hit that cap in the middle of a busy day.

A worked example helps. Take a Tuesday with lots of email. You dictate replies in Mail, notes in Notes, and quick messages in Slack. On Superwhisper's free tier, the local model handles all of that without a plan change, as long as you stay within the two non-AI modes. If you want the AI modes, you need Pro. On Typeless, the free tier covers the first part of the day, then the word cap decides what happens next.

The common mistake is judging a free plan by the first hour. Run each one for a full workday. Count how often you wait on a limit or switch back to the keyboard. That count tells you more than any feature list.

Where your voice goes: local models or cloud servers

This is the biggest difference between the two. Superwhisper's privacy policy, last updated October 8, 2026, says: 'When you use a local model, your audio and transcripts never leave your device.' Its cloud models process audio and transcripts in memory and do not store them after the request completes.

Typeless's privacy policy, last updated October 13, 2026, takes a different route. It says voice audio and limited context are processed in real time on its cloud servers. It also says audio is immediately discarded once the transcription result is returned to your local device. The policy does not describe on-device transcription.

Typeless's homepage makes stronger claims. It says 'Zero cloud data retention', 'Never trained on your data', and 'No audio stored in the cloud'. It also says Typeless is ISO 27001 certified, GDPR compliant, and HIPAA compliant. The privacy policy names no certifications. Treat the homepage claims as marketing until you can read the certificates or a trust page.

You can test the local claim yourself. Pick a local model in Superwhisper's settings. Turn Wi-Fi off from the menu bar, then dictate a sentence into Notes. If the text appears, the model runs on your Mac. Then try the same test in Typeless. If Typeless stops working offline, that fits its cloud policy. Run the test on your own machine, because results depend on your model and your hardware.

Privacy claims to check before you trust them

Superwhisper's privacy policy says server logs are kept for 30 days, and service usage metadata for up to 3 months. It also says its data is not used for training AI models. The SOC 2 Type II claim sits on its homepage and applies to Enterprise. The privacy policy names no certifications, so you cannot confirm that claim from the policy alone.

Typeless's policy says data sent to third-party LLM providers, including Microsoft and Google, is never used to train those services. That is a promise about the providers. If your work is sensitive, read the provider terms too, not just Typeless's page.

If you handle client files, medical notes, or unreleased code, write three questions down. Where does the audio go? How long is it kept? Who can see it? Then match each answer to a specific paragraph in the policy, not to a homepage line. A good answer names a time limit and a place.

Accuracy and reliability: what users say

Reviews are thin on both sides, so read them as hints. Superwhisper has a 4.9 rating on Product Hunt from 22 reviews. One reviewer wrote that 'the transcription accuracy blows me away every time.' Another wrote that 'it is not 100% reliable.' Both are opinions, and the sample is small.

Typeless has a 3.5 TrustScore on Trustpilot from only 3 reviews. Trustpilot notes that the profile is unclaimed and that the company has not invited reviews. That tells you little about accuracy. It does tell you the company is not collecting feedback on that site.

Accuracy depends on your voice, your microphone, and your vocabulary. Run your own test. Read the same paragraph aloud in each app. Compare the output with the original text. Count the errors on names, numbers, and product terms. Those are the words that cost you time in real work.

The dictation accuracy checker can help you score a sample. Use the same sample for both apps so the comparison stays fair.

Platforms: Mac, Windows, phones and Linux

Superwhisper's docs list macOS, Windows, iOS, and Android as available. The docs mark Linux as coming soon, but the homepage lists Linux. Meetings is macOS only. It records calls in Zoom, Google Meet, Microsoft Teams, or Slack without a bot.

Typeless lists desktop macOS, Windows, and Linux, plus iOS and Android. On paper, that is the wider list. For a Mac-only workflow, though, the list matters less than how the app behaves on your machine.

Intel Macs need extra care. Superwhisper's docs say Intel Macs work best with cloud models. Offline models only run well on Apple Silicon Macs. To check your chip, choose Apple menu > About This Mac. Apple silicon shows as the chip, and Intel shows as the processor.

Features that matter for daily Mac work

Superwhisper's docs list intelligent and custom modes, a 'Write for me' mode, and context-aware AI that uses selected text and the clipboard. It supports over 100 languages with translation to English. It also connects to coding agents such as Claude Code, Cursor, and Codex.

Pro adds cloud voice models, file transcription, unlimited modes, AI-powered modes, context awareness, realtime transcription, system audio recording, full Meetings on macOS, bring-your-own API keys, and priority support. If you want to use your own API keys, Pro is the plan that allows it.

Typeless's homepage describes removing filler words and repeats, keeping the final wording after self-corrections, a personal dictionary, and style that adapts to each app. It lists 100+ languages with automatic detection, real-time translation, and voice commands that edit selected text.

Typeless Pro adds unlimited words, enhanced accuracy, priority access during high demand, cloud sync, team management, centralized billing, and team usage analytics. If you buy for a team, those admin tools are the reason to look at Typeless first.

Choose by the kind of control you want. Superwhisper's modes let you decide the output style. Typeless's cleanup decides much of it for you. Neither is wrong. A developer who wants a fixed commit message format may prefer modes. A writer who wants clean prose may prefer the cleanup.

How much faster is speaking than typing?

A 2016 arXiv study by Ruan, Wobbrock, Liou, Ng and Landay found speech input was 2.93 times faster than keyboard entry for English. The rates were 153 vs. 52 WPM. For Mandarin, speech was 2.87 times faster, at 123 vs. 43 WPM. The test used touchscreen phones and short messages.

That is a phone study, not a Mac study. The gap still matters. If you write a lot of email or notes, speaking can save real time. You can measure your own speed with the typing speed test and the speaking vs typing test.

Speed is not the whole story. Dictation also means fixes. The mistake is counting only the words per minute and forgetting the time you spend correcting a name. Budget that time in your test.

Edge cases: names, quotes, code and mixed languages

Names and product terms are where dictation breaks first. Superwhisper lets you add a custom vocabulary and text replacements. Typeless offers a personal dictionary. Load your client names, product names, and team handles into whichever app you pick before you judge accuracy.

Cleanup can change a quote. Typeless removes repeats and filler words, which helps with speech but can change exact wording. If a quote must match word for word, check the output against your recording before you send it.

Mixed languages need a test too. Superwhisper lists over 100 languages with translation to English. Typeless lists 100+ languages with automatic detection and real-time translation. If you switch languages mid-sentence, dictate one sentence like that in each app and compare the results.

Code is its own case. Superwhisper's docs connect modes to coding agents such as Claude Code, Cursor, and Codex. The Typeless pages we checked do not describe a coding connection, so check its docs before you plan a coding workflow around it.

Who should pick Superwhisper

• You want unlimited dictation on a free plan, with local models on your Mac.
• You handle sensitive text and want the audio to stay on the device.
• You want custom modes, text replacements, and file transcription.
• You code and want modes that connect to tools such as Claude Code, Cursor, or Codex.

Pick Superwhisper if control matters more than polish. Expect to spend a little time in settings. Local models run best on Apple silicon, so check your chip before you rely on offline work.

Who should pick Typeless

• You want filler words and self-corrections cleaned up with little setup.
• You work across Mac, Windows, and phones and want the same app on each.
• You run a team and need management, centralized billing, and usage analytics.
• You accept cloud processing for your voice and want to test the free tier first.

Pick Typeless if you want the output to read clean without tuning. Check the word cap on the free plan first. If your work is sensitive, read its privacy policy before you send client text through it.

Where Lazur fits if you want more than dictation

Lazur is a voice assistant for the Mac that lives in the notch. Hold Option and Command to ask about anything on screen, and Lazur answers out loud. Hold fn to dictate clean, formatted text at the cursor in any app. Dictation is included, but it is not the headline. It also has snippets, a personal dictionary and clipboard slots.

Lazur looks at the screen only while you hold the shortcut. Ask where something is, and it draws a circle and a label around it. Ask how to do something, and it guides you one click at a time, waiting for each click. It does not click buttons, edit files, or operate other apps for you. It is not a plugin or add-in for any app.

Lazur runs on macOS, on Apple silicon and Intel. Windows is planned, not available. Lazur says audio and the screenshot go to AI providers to answer, and are not stored. It also says content is never used to train AI models. Lazur keeps a memory of what you asked and did, on the Mac.

Lazur's pricing starts with a 3-day Pro trial, and no credit card is needed. Pro is $18/month, or $15/month billed annually, with 150 AI actions. Power is $120/month, or $96/month annually, with 1,000 AI actions. Dictation is unlimited on both.

If you only need dictation, Superwhisper's free tier or Typeless may cost you less than Lazur's Pro plan. Lazur makes more sense when you also want the assistant layer. Be honest about which job you have.

Setting up each app before you judge it

1. Give the app microphone access in System Settings > Privacy & Security > Microphone.
2. If an app asks for accessibility access, allow it in System Settings > Privacy & Security > Accessibility.
3. Pick your speech model. For offline work in Superwhisper, choose a local model and confirm your chip first.
4. Add your names, product terms, and team handles to the custom vocabulary or personal dictionary.
5. Dictate the same short email in each app, in the apps you use most.
6. Compare the output with your original words, and count the fixes.
7. Check the privacy page for the model or mode you used. Is it local or cloud?
8. Only then compare plans, and pick the billing period you would actually use.

Give each app the same test on the same day. Keep short notes as you go. You will remember the annoying fix more than the good result, so write it down while it is fresh.

Mistakes people make when comparing dictation apps

• Comparing headline prices and skipping the free tier limits.
• Testing in a quiet room, then judging accuracy in a noisy office.
• Treating a homepage claim as the privacy policy. Read the policy text.
• Ignoring Intel. Offline models behave differently on Intel Macs.
• Buying a yearly plan before the free tier shows the app fits your work.
• Reading a rating from a small sample as a verdict.

Also watch for the Superwhisper homepage price label. It shows $849 on one line, which looks like a rendering error for $8.49. Use the billing plans page instead, and check the date on any figure you copy.

What to do next

Start with the question that decides the choice for you. If audio privacy decides it, read Superwhisper's local model rules and Typeless's privacy policy side by side. If price decides it, compare the free tiers on your real work. If you want an assistant that sees your screen, add Lazur to the test.

For a closer look at each pairing, read Lazur vs Typeless on Mac in 2026 or Superwhisper alternatives for Mac in 2026. You can see Lazur's plans before you install anything.

Pick one app, use it for a full week, and keep a list of what annoyed you. That list usually makes the decision for you.