Wispr Flow Review 2026: Pricing, Privacy, and Verdict
Wispr Flow is the most heavily funded product in voice dictation, and it shows in the polish. It runs on Mac, Windows, iPhone and Android, it covers more than 100 languages, and the writing model behind it is good at working out how you want to sound in a given app without being told.
We build Kana, which competes with it on the Mac. Read this knowing that, and hold us to the parts you can check on Wispr’s own site.
Short answer: Wispr Flow is the right pick if you dictate across more than one operating system, if you want a phone app, or if you want the most polished writing model in the category. Recognition happens on Wispr’s servers, and if that is a problem for your work, no privacy setting changes the fact that the audio leaves your machine.
What Wispr Flow actually is
Hold a shortcut, speak, and formatted text lands wherever your cursor is. That part matches every other dictation app on this site. What Wispr adds is a writing layer on top: it strips filler words, punctuates, and adapts the register depending on where you are typing, so a Slack message and a board update do not come out sounding the same.
Its own homepage puts the keyboard at 45 words a minute and Flow at 220, which it rounds to “4x faster than typing”. That is Wispr’s number, measured by Wispr. We have not tried to reproduce it.
The app also learns names and terms you use, which it describes as Flow learning your vocabulary automatically rather than as a dictionary you fill in by hand. There is a separate product, Notetaker, for meeting notes, and Wispr’s pricing page says it is Mac only for now.
Source: Wispr Flow (wisprflow.ai)
Pricing in 2026
There is a free plan at $0. It gives you dictation across all your apps in 100+ languages, with weekly caps that differ by device: 2,000 words a week on desktop, 1,000 a week on iPhone, and unlimited on Android. Notetaker is included with a weekly limit.
Pro is $15 per user a month, or $12 per user a month billed annually, which Wispr marks as a 20% saving. Enterprise pricing is quoted by their sales team.
The desktop cap is the number that matters. Two thousand words a week is roughly four decent emails a day for a working week. If dictation becomes how you write, you will hit it in the first month and the free plan stops being a plan.
Where your voice goes
This is where the two products actually differ, so it is worth being precise rather than dramatic.
Wispr’s privacy page says that with Privacy Mode turned on, they keep nothing: no audio, no transcripts, no edits, and none of it is used by Wispr or a third party to evaluate, train or improve AI models. Turning on Privacy Mode and disabling Private Cloud Sync gives you what they call zero data retention. Without Privacy Mode, they say your data may be used to improve Flow’s features and models, and that it is never sold or shared. They also list SOC 2 Type II, ISO 27001 and HIPAA certifications.
That is a serious posture, and better than most cloud dictation apps offer. It is also a promise about storage rather than about transmission. Your audio is still recognised on Wispr’s infrastructure, so dictation needs a working connection, and the recording exists on someone else’s machine for as long as it takes to turn into text.
Kana makes zero network calls when you dictate. Not “we delete it quickly”, not “we do not train on it”: the app does not open a socket. That is checked in our CI rather than written on a page. Whether that difference is worth anything to you depends entirely on what you dictate.
| Wispr Flow | Kana | |
|---|---|---|
| Recognition runs | On Wispr’s servers | On your Mac |
| Works with no network | No | Yes |
| Mac, Windows, iPhone, Android | Yes | Mac only, Apple Silicon |
| Languages | 100+ | 26 |
| Trial | 2,000 words a week on desktop | 14 days, no card |
| Paid | $15 a month, $12 annual | $4.90 a month or $49 a year |
| Meeting notes | Yes, Notetaker | No |
| Learns your vocabulary automatically | Yes | Custom vocabulary you enter |
| Account required | Yes | No |
What Wispr Flow does better than Kana
Three things, and the first one ends most comparisons before they start.
It runs where Kana does not. Windows, iPhone, Android. Kana is macOS on Apple Silicon and will stay that way, so if your day crosses two operating systems, Kana solves half your problem at best and you should stop reading here.
Its writing model is better at tone. Wispr infers register from context, so the same spoken sentence comes out shaped differently in a code review than in an email. Kana applies no tone of its own unless you turn on AI rewriting and write the instruction yourself. For people who do not want to think about it, Wispr’s default is the better default.
It covers 100+ languages against Kana’s 26. Kana runs Parakeet on the Neural Engine for English and 24 other European languages, and Apple’s SpeechAnalyzer for Japanese. That is the whole list. If you dictate in Hindi, Arabic or Thai, Kana has nothing for you.
Source: Wispr Flow (wisprflow.ai)
Where a local Mac app wins
Kana does one job and stops. Hold a key, speak, release, the words land in the frontmost app. English and 24 other European languages run on Parakeet on the Neural Engine, Japanese runs on a second engine built for it, and none of it needs a connection.
Kana is $9.99 once, and that buys unlimited dictation with no per-word cap, plus transcribing audio and video files with SRT, VTT and Markdown export. AI rewriting is the one subscription: $4.90 a month or $49 a year, covering three Macs. There is no account to make and nothing in the Dock.
The app bundle is 2.4 MB. The speech model is a separate download of about 480 MB, so the install is not 2.4 MB in total, but there is no background service running while you are not dictating.
We have run no head-to-head benchmark against Wispr Flow, so there is no accuracy table here and no latency chart. Both would be marketing dressed as measurement.
If you want the two lined up feature by feature, the Kana and Wispr Flow comparison does that, including the rows where Wispr wins.
Who should use Wispr Flow
- You dictate on Windows, an iPhone or an Android phone as well as a Mac
- You want the output to sound right without configuring anything
- You dictate in a language outside Europe and Japanese
- You want meeting notes and dictation from one vendor
- Your employer is happy with a SOC 2 and ISO 27001 vendor holding the audio path
Who should use Kana instead
- You are on Apple Silicon and nowhere else
- Your work makes sending recordings to a third party awkward or forbidden
- You dictate on planes, trains and bad hotel wifi
- You want dictation to be $9.99 once rather than $15 a month
- You transcribe files and would rather not pay for that separately
FAQ
Is Wispr Flow free?
There is a free plan. It gives you 2,000 words a week on desktop, 1,000 a week on iPhone and unlimited on Android, in 100+ languages. Pro removes the caps at $15 a month, or $12 a month on annual billing.
Does Wispr Flow work offline?
Wispr does not claim on-device recognition on its site, and the app is built around a cloud recognition path, so you should assume dictation needs a connection.
Does Wispr Flow store my voice?
By default your data may be used to improve Flow’s features and models, which Wispr says is never sold or shared. Privacy Mode changes that: with it on, Wispr says it keeps no audio, no transcripts and no edits, and disabling Private Cloud Sync as well gives zero data retention.
What is the best offline Wispr Flow alternative on a Mac?
Any dictation app that runs the model locally, ours included. The filter is what you would be giving up: Windows, phones, 100+ languages, and meeting notes. If you need one of those, staying with Wispr is the cheaper decision even when the invoice is bigger.