Wispr Flow or your Mac's free voice typing: which to keep

Pick Wispr Flow if you dictate text that carries names, jargon or a full argument and you can be online; pick the dictation built into macOS if your speaking runs to a few short notes and the audio has to stay on the machine. The deciding test is edit count: dictate one real message into each and count what you fix before sending. Wispr Flow has been our main text input since 2024, and Apple's tool we have never measured on our own machines; every fact about it below comes from Apple's support pages, and every Flow figure from Wispr's own.

Run one real message through both

Take a message you actually need to send, at least four sentences long, with one name in it that autocorrect usually mangles. Dictate it into Apple's tool first: press the Microphone key or the keyboard shortcut, or choose Edit > Start Dictation, speak, and count the edits before you would hit send. Then try the same message in Flow. Wispr publishes this ten-minute test on its own comparison page, and it is the right one, because the difference between the two is not speed, it is what you fix afterwards.

Two things show up in that single message. Apple's dictation gives you roughly the right commas and periods, and it also gives you your filler words, so a spoken "um" lands in the text. Flow removes filler words and false starts as you speak, adds punctuation and capitalization, and keeps a list of your names and product terms that you can correct once and never again. The names are where Apple's tool stops: its guides describe voice commands and optional sharing of your audio recordings for training Apple's models, and no dictionary that learns your jargon.

Price the switch before you make it

Apple's dictation costs nothing and ships with the device, on Mac, iPhone and iPad. Wispr Flow's pricing page, read in United States dollars, lists Free at 0, Pro at 15 dollars per user, per month, or 12 billed annually, Growth at 23 monthly or 18 annually, and Enterprise on quote. Pro is the tier that matters for a solo switch: unlimited dictations, the shared dictionary, team workspace, centralized billing and priority support. Growth adds single sign-on, admin controls and unlimited notetaker history.

The free plan is a permanent tier, not a trial, and its limits are published in the Help Center: a 2,000-word weekly allowance on Mac or Windows with a 5,000-word maximum, resetting Sunday at 12 a.m. PT. On iPhone it is 550 words allowance and 750 maximum, on Android 2,500 and 3,000, and the mobile allowances draw from one pooled count. At the allowance Flow slows down; at the maximum it blocks until the reset or an upgrade, and limits apply at your next press, never during an accepted dictation. An account opened from the link on this page starts with one month free on the paid tier, a perk we negotiated with the vendor and verified on 22 September 2026; no public page announces it.

Read the 30 seconds for what it is

The number that fills the support forums is real, and Apple's own troubleshooting page states it plainly: no speech detected for 30 seconds and dictation stops on its own. The same page says you can dictate text of any length without a timeout, and that you can also turn the feature off manually. So the 30 seconds is a silence cutoff, not a cap on how much you can say, and Wispr's comparison page adds what Apple does not offer: no setting changes the silence stop.

Flow starts and stops differently. You hold or press to start, and it stops when you release the key, so a pause for thought costs you nothing but a keypress. That is the whole practical difference for long messages: with Apple's tool, a sentence that needs thinking through is a race against the silence timer, and with Flow it is not. On Mac, one Apple-side conflict is worth knowing before you test: when Voice Control is on, standard dictation is not available.

Test a real week before you pay

A week is the right window, and the free plan fits one. Our own dashboard capture, taken 25 September 2026 on a production account, reads 1,145,134 words dictated at an average of 113 words per minute, across 144 distinct applications, with a 27-day streak running. The typing reference is the Aalto University study from CHI 2018, with an average of 52 words per minute. Between those two rates, the same text takes more than twice as long at the keyboard as it does spoken aloud, which is the whole commercial case for either tool.

Two honest caveats go with that comparison. The 113 words per minute is a dictated counter, and it says nothing about the re-reading that turns a transcript into finished prose. And 52 is a study average, not this writer's keyboard speed. What the week should test instead is accuracy on your own material: your clients, your product names, your jargon. Set the language explicitly rather than relying on auto-detect, because the Wispr Flow product page says a wrongly selected language confuses the transcription and that selecting just the language you are speaking gives noticeably better accuracy.

Watch what breaks when the internet drops

Flow is cloud-only, and Wispr says so on its own comparison page: transcription runs in the cloud and needs a connection on every platform, with no offline mode anywhere. Starting offline by hold-to-talk shows "No internet" in the app, per the Help Center. On a plane or in a basement office, Apple's tool on a supported model keeps working, because Apple processes general-text dictation on the device and sends nothing to Siri servers.

The Flow side has published edges worth knowing before you commit. Saved audio expires after 14 days, except recordings still awaiting upload, and the inline retry needs a failed or empty recording of at least five seconds. If your data setting is "never store," history is disabled and no audio is kept for retry, recovery or download. AirPods and Bluetooth earbuds are the one hardware setup Wispr tells you to avoid, since they hurt accuracy. And there is no iPad app, so the tablet stays on Apple's tool. On the Apple side, the privacy trade is optional and explicit: you can share audio recordings of your Siri and dictation interactions so Apple can store them, Apple may review a sample, and the toggle lives in Privacy & Security > Analytics & Improvements.

Stay on Apple's dictation if your week looks like this

The switch is wrong for a real set of people, and our tool sheet says which. If your dictation is short notes, searches and quick messages, the free, built-in tool is enough and the 30-second rule never bites. If you work offline or your constraint is audio that never leaves the machine, Apple's on-device processing is the architecture you want. If your writing is mostly code and symbols, neither tool helps and the keyboard stays. And if you speak from an iPad, Flow has nothing to run.

We also ran a fuller comparison against a local, one-time-purchase alternative in VoiceInK vs Wispr Flow, and the wider question of what dictation is still bad at is in dictation that finally works. Wispr Flow is an affiliate: we earn a commission if you sign up through our link, on top of the free month that comes with it.