Skip to content

Dictation software

Wispr Flow: Price, Privacy and What Users Say

Wispr Flow is a cloud dictation app that turns what you say into edited text in any app on Mac, Windows, iPhone and Android. It is free with weekly word limits, and Pro costs $15 per user per month, or $12 billed yearly. Every dictation is transcribed on Wispr's servers, so it needs an internet connection, and its data settings start switched on.

Updated October 27, 202616 min read106 cited facts, each traced to its source

15US dollars per user per month

On Wispr Flow's pricing page, the Pro plan for individuals and teams costs $15 per user per month under the Monthly setting of the price switch.

Wispr Flow, Wispr Flow (vendor website), 2026

18,110ratings

Wispr Flow had 18,110 ratings on the United States App Store on October 26, 2026.

How it is calculated

1,000,000downloads or more

Wispr Flow showed 1M+ downloads on Google Play (at least 1,000,000), rated 4.6 stars from 13,116 ratings, on October 26, 2026.

How it is calculated

280million US dollars

Wispr announced $280 million in Series B funding at a $2 billion valuation, led by Menlo Ventures, with Notable Capital as a returning investor.

Tanay Kothari, Wispr Flow, 2026

Wispr Flow is an AI dictation app from Wispr AI that turns speech into clean, punctuated text in whatever app you are typing in, on Mac, Windows, iPhone and Android. It is free to start, with a weekly word allowance, and its Pro plan costs $15 per user per month 1, or $12 a month when billed yearly 2. Its users like it: the US App Store showed 18,110 ratings 3 averaging 4.85 stars 4, and Google Play showed at least 1 million downloads 5.

The trade-off is the cloud. Wispr says transcription always happens on its servers 6, so the app needs an internet connection 7, and two data settings, model training and cloud storage of your dictations, start switched on unless your plan or organization restricts them 8. Prices, limits and policies below come from Wispr's own pages and app store listings, read between October 16 and October 27, 2026. For the wider field, see our overview of dictation software and the comparison of AI dictation apps.

Wispr Flow at a glanceDetail
MakerWispr AI, Inc.
What it doesCloud speech to text with AI cleanup, in any app
PlatformsMac, Windows, iPhone, Android 9
Not supportediPad, Linux, Chromebooks 10
LanguagesMore than 100 11
Free plan$0, with weekly word limits 12
Pro$15 a month 1, or $12 a month billed yearly 2
Works offlineNo 7
Funding$280 million Series B at a $2 billion valuation 13

Wispr Flow is a cloud AI dictation app that rewrites what you say

Wispr Flow does more than transcribe. It is an AI dictation app: a speech recognizer turns your voice into words, then a language model edits them. The home page says it removes filler words, adds punctuation and formats your writing as you dictate 14, and that it works anywhere you can type, with no plugins 15. You can ramble and pause, and the cleanup step removes the filler words for you.

On a computer you hold a shortcut, speak and release. The defaults are Fn on Mac and Ctrl + Win on Windows 16. Desktop dictation inserts the finished text after you stop, rather than showing a live word-by-word transcript 17. That design is what lets the app clean up a whole passage before you see it, and it is also why a recognition error only shows once the paragraph lands. On iPhone, Flow is a keyboard you switch to; on Android, it is a floating bubble.

Three features go beyond plain dictation:

  • Context awareness reads limited text from the app you are dictating into, so names and jargon come out right 18. It is on by default 19.
  • Dictionary and snippets learn your names and terms, even on the free plan 20. This is Wispr's form of custom vocabulary.
  • Command Mode lets you edit text by voice 21.

How a single dictation travels

The engine underneath has changed over time. In mid-2025, 9to5Mac reported that Flow's intelligence was based on a mix of Meta's Llama and OpenAI's Whisper models 22, which is why people ask whether Whisper is the same as Wispr Flow. It is not: Whisper is an open speech model, and Wispr Flow is a commercial app. In November 2025 the company claimed an error rate of around 10%, against 27% for OpenAI's Whisper and 47% for Apple's native transcription 23. Those are vendor numbers with no published test set, so treat them as claims, not as an independent word error rate benchmark. Our guide to AI dictation cleanup explains what the editing step changes.

The company did not start with software. Its CEO set out to build a device that would let people type by silently mouthing words, and its early funding went to that hardware 24. Flow was the software interface of that device. It reached the Mac in October 2024, Windows in March 2025 and iPhone in June 2025 25, and Android on February 23, 2026 26.

Free, Pro and Enterprise plans differ by word limits and features

The free plan is real, but it is capped by words per week, and the cap is lower on phones than on computers. Pro removes the cap. Growth and Enterprise add team administration and compliance controls. If you dictate past the weekly limits described below, you will need Pro.

PlanMonthly billingAnnual billingDictation limitWhat it adds
Free$0 12$0Weekly cap, by platformAny app, 100+ languages 12, dictionary, snippets 20
Pro$15 per user 1$12 per user 2Unlimited 27Shared dictionary and snippets
Growth$23 per user$18 per user 28Not stated on the pricing pageBasic admin controls, model training off and locked 29
EnterpriseCustomCustomNot stated on the pricing pageModel training off and locked 29

With unlimited Notetaker, Wispr's meeting-notes product, Growth costs $26 per user per month billed yearly or $33 billed monthly 28. Billed yearly, one Pro seat comes to about $144 a year 30, against $180 at the monthly price 31. Students and educators with an education email get 50% off Pro, and non-profits get a discount 32. On iPhone, the US App Store sells Pro at $15.00 a month or $143.99 a year, a Student plan at $7.49 a month, and a weekly Pro plan at $4.49 33.

The free word limits by device

The pricing page advertises 2,000 words a week on desktop 34. Wispr's help center gives the full picture: on Mac and Windows a limit message appears after 2,000 words, but new dictations are blocked only above 5,000 words a week 35. The desktop count combines desktop and mobile dictation 36. On Android, dictation slows after 2,500 words and is capped at 3,000 37. On iPhone, it slows after 550 words and is blocked at 750 38.

To put 2,000 words in perspective: averaged over a week it is about 286 words a day 39. The free tier suits trying the app or light use; a writer drafting a long document would exhaust it quickly, which is the point of a free tier. Our guide to what AI dictation costs explains why AI dictation apps charge a subscription: every cloud dictation costs the maker server time.

Cancelling and refunds

Cancelling does not delete your account or data, paid access runs to the end of the billing period, and there is no automatic prorated refund 40. Subscriptions bought through the App Store or Google Play are cancelled in that store. For a fuller answer, see whether Wispr Flow is free and whether it is worth paying for.

Your audio is processed in the cloud, so its privacy terms matter

Wispr Flow is a cloud dictation app with no on-device mode 6. Your voice leaves your device every time you dictate, which makes its retention and training settings the main privacy question. The short answer: the company encrypts data and lets you switch off storage and training, but on Free and Pro both start switched on, so check them on day one.

What happens to your audio

Audio is captured on your device only after you press the shortcut, then sent to Wispr's servers over TLS 1.2 or later 41. The privacy page says audio is deleted as soon as it is transcribed, unless you have turned on Dictation cloud storage 42. The help center says that cloud storage switch keeps transcripts, audio and history on Wispr's servers for cross-device access, and that it defaults to on unless restricted 8. Anything stored is encrypted with AES-256 43.

Wispr may also send data to outside AI providers. Its privacy policy says third-party AI or language model providers may process customer content to provide the service 44, that they do not use it to train their models, and that it is generally deleted within 30 days 45.

The settings that control your data

SettingWhat it doesDefaultSource
Improve the model for everyoneLets Wispr use your audio, transcripts and edits to train its modelsOn for Free and Pro; off and locked on Enterprise 46
Dictation cloud storageKeeps dictation audio, transcripts and history on Wispr's serversOn unless restricted 8
Context awarenessReads limited text from the app you dictate intoOn 19
Local data storageKeeps history on your computerYour choice: keep it, delete it after 24 hours, or never save 47

Turn off Improve the model for everyone and, Wispr says, nothing you say is used to train its models 48. On Growth and Enterprise, training is off by default and locked 29. Wispr says that turning off context awareness means the app transcribes without reading anything on your screen 49. When it is on, standard password fields are never read 50 and banking apps are skipped 51, but Wispr warns that custom or web-based password fields may be read 52. Our question on whether AI dictation reads your screen compares how other apps handle this.

Wispr says it never sells your data and that its business is selling software 53. Its App Store privacy label does list identifiers and usage data as data that may be used to track you across other companies' apps and websites 54. The label names only identifiers and usage data in that tracking category.

HIPAA and audits

Any user can sign a business associate agreement to enable HIPAA support for dictation, on any plan, and accepting it turns off Notetaker and Scratchpad 55. A signed BAA also locks both data-sharing controls off 56. Wispr's security FAQ notes that its SOC 2 and ISO 27001 audits covered dictation, and that Notetaker launched after those audit periods and was not assessed 57. A BAA is a contract, not a guarantee that your use is compliant; our guide to HIPAA-compliant dictation explains what else a practice needs. For retention periods across apps, see the guide to voice data retention, and for the full safety answer, whether Wispr Flow is safe.

It runs on Mac, Windows, iPhone and Android

Wispr Flow runs on the four big platforms, but only on fairly recent versions, and not on iPad, Linux, Chromebooks or Windows on ARM. It always needs a connection.

PlatformMinimum systemHow you dictateWorth knowing
MacmacOS 12 Monterey or later, Apple silicon or Intel 58Hold Fn, speak, releaseNeeds Accessibility access to paste 59
WindowsWindows 10 or 11, x64 processor 58Hold Ctrl + WinARM and Snapdragon PCs are unsupported 60
iPhoneiOS 18.3 or later, iPhone only 61Switch to the Flow keyboardNeeds Full Access for the keyboard
AndroidAndroid 13 or newer, phone with an accelerometer 61Tap or hold the Flow BubbleSmartphone required

Both desktop apps recommend 8 GB of memory 58. On Windows, the lack of ARM support rules out Copilot+ PCs with Snapdragon chips; our page on dictation on Windows on ARM lists what does run there. On iPhone, Flow is a third-party keyboard with Full Access, which is worth understanding before you enable it: see whether third-party keyboards can see what you type. Platform questions are answered in detail in Does Wispr Flow work on Windows, Android and Linux?

There is no offline mode on any plan: an internet connection is required for transcription 7. If you dictate on planes, in hospitals with locked-down networks, or simply want audio to stay on your machine, you need an app with on-device speech recognition. Our question on whether Wispr Flow works offline lists local alternatives.

When Wispr Flow is not working

Most problems are a missing permission, a stolen shortcut or a microphone the system cannot hear. Work through these in order (Wispr Flow help center, October 2026):

  1. Recover the text. If Flow heard you but nothing appeared, press Ctrl + Cmd + V on Mac or Alt + Shift + Z on Windows to paste the last transcript 62.
  2. Check Accessibility on Mac. Open System Settings > Privacy & Security > Accessibility and enable Wispr Flow, or switch it off and on, then quit and reopen Flow. Windows does not need this permission 59.
  3. Test the microphone outside Flow. Record a short clip in Voice Memos or Sound Recorder; silence there points to the microphone or system setup, not Flow 63.
  4. Pick the microphone by name in Settings > General > Microphone instead of auto-detect, then restart Flow.
  5. Check the connection. Without internet, Flow cannot transcribe at all 7.

Some slowness is the app itself: a How-To Geek reviewer on Windows reported that Wispr Flow used 800 MB of RAM even when idle and took 8 to 10 seconds to start 64. That is one reviewer's measurement, not a specification. More fixes are in Why is Wispr Flow not working?

Its ratings and installs show a large, satisfied user base

Wispr Flow has far more US App Store ratings than the three rivals charted below, and its average is high in every country we measure.

App Store countryRatingsAverage stars
United States18,110 34.85 4
United Kingdom2,703 654.80 66
Germany1,865 674.85 68
France776 694.83 70
Brazil624 714.91 72
Japan177 734.75 74

The US store holds about 75% of the ratings across these six countries 75. The company's own account of where users were in June 2025 is a different mix: 40% in the US, 30% in Europe and 30% elsewhere 76. Ratings are not users, so the two figures cannot be compared directly. India is missing from our rating sample, yet the CEO says it is the app's second-largest market by both users and paying subscribers 77. On October 26, 2026, the iPhone app ranked number 48 among top-grossing Productivity apps in the US 78, 28 in the UK 79 and 24 in India 80.

Willow Voice had 1,073 US ratings 81, Superwhisper 846 82 and Aqua Voice 99 83. On Google Play, Wispr Flow's 1M+ band 5 compares with at least 500,000 for Typeless 84. Usage figures from the company point the same way: people have written more than 60 billion words with Flow 85, and by November 2025 it had reached 270 of the Fortune 500 86. Developers talk about it too: "wispr flow" appeared in 137 Hacker News stories and comments in the 12 full months before September 2026 87.

What users praise and what they criticize

Wispr Flow is popular for reasons the product makes plain: it works in any app without plugins, its cleanup removes the filler words that make raw dictation hard to use 14. The common complaints are just as concrete. It needs a connection for every dictation. The free mobile allowance is small. It does not run on iPad or Linux. And because data settings start switched on, privacy-minded users have to go and change them.

Wispr has raised about $361 million and was valued at $2 billion

Wispr's funding has grown quickly since Flow launched. Adding the August 2026 Series B to the earlier total gives about $361 million in disclosed funding 88.

RoundDateAmountLed by
New investment at Flow's launchSeptember 2024$12 million 89Matt Kraning
Series AJune 2025$30 million 90Menlo Ventures
Series A extensionNovember 2025$25 million 91Notable Capital and Flight Fund
Series BAugust 2026$280 million 13Menlo Ventures

By November 2025 the company had raised $81 million in total, and sources put its valuation at $700 million 92. The August 2026 Series B valued it at $2 billion 13. Wispr makes money from subscriptions, not from data, by its own account 53. See How much funding has Wispr Flow raised? for the investors and what the money is for.

Who it suits and when another app fits better

Wispr Flow suits people who write a lot, across several devices, and are comfortable with cloud processing: managers clearing email, founders writing memos, developers prompting AI tools, and teams that want a shared dictionary and admin controls. It is a weaker fit if you need offline use, keep sensitive material off third-party servers, work on an iPad, Linux or an ARM PC, or dictate only occasionally, where the free tools built into your system may be enough.

AppWhere speech is processedPlatformsFree optionPaid entry price
Wispr FlowCloud onlyMac, Windows, iPhone, AndroidWeekly word cap$15 a month 1
SuperwhisperOn device with local Whisper models 93Mac, Windows, iPhone, Android 94Unlimited local Whisper models 93$8.49 a month 95
Aqua VoiceCloud 96Mac, Windows, iOS, Android 971,000 free words 98$12 a month 99
Willow VoiceCloud 100Mac, Windows, iPhone 101Basic plan with a weaker model 102See the Willow guide
Apple DictationOn device for general dictation on Mac 103Apple devicesFree, built inNone
Windows voice typingOnline, Azure Speech 104WindowsFree, built inNone

Against its closest rivals, Wispr Flow's advantages are reach and polish: four platforms, and the most US App Store ratings of the apps compared here. Superwhisper costs less and can keep speech on your Mac; Aqua Voice is the nearest cloud alternative; Willow Voice competes on the same AI cleanup; and VoiceInk is open source 105 and local 106. Head-to-head answers are in Wispr Flow vs Superwhisper, Aqua Voice vs Wispr Flow, Willow Voice vs Wispr Flow and VoiceInk vs Wispr Flow.

Built-in dictation is the honest baseline. Apple Dictation is free and processes general dictation on the Mac itself 103, and Windows voice typing is free but online. Neither rewrites your rambling into finished prose the way Wispr Flow does, which is the feature you pay for. If you mostly dictate short messages, start with the built-in tool; see Apple Dictation vs Wispr Flow and the guide to free dictation software.

If you want dictation without the cloud, Yap, which we make, runs speech recognition on your Mac or Windows PC; we compare the two fairly on our Wispr Flow alternative page. For a broader shortlist by device, see the best dictation apps for Mac, the best dictation software for Windows, the best dictation apps for iPhone and Android and our ranking of dictation software by need.

Questions about Wispr Flow

Evidence

Every number on this page, with where it comes from. Open a fact to see the exact words in its source.

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.

The data behind these numbers

Questions about this

More in Dictation software

Referenced by