
by Wispr AI
AI dictation for Mac, Windows, iOS and Android that cleans up speech and formats it for the app you are typing in, plus a meeting Notetaker.
Wispr Flow suits people who write all day and want clean, punctuated text from speech in any app across Mac, Windows, iOS and Android. It is free to start with a weekly word limit, and larger teams can lock model training off. It always transcribes in the cloud, so it needs an internet connection.
Reformats speech for whichever app is in front of you, now on Wispr's own Canto model.
Wispr says it is already training Canto's successor at more than ten times the scale, per its release notes of 17 September 2026.
Main trade-off
Transcription always runs in the cloud, so it cannot work offline.
| Plan | Monthly price | What it includes |
|---|---|---|
| Free | Free | Dictation on every supported platform, in 100+ languages, A weekly word limit, A personal dictionary and snippets, Limited Notetaker access |
| Pro | $15/mo | Everything in Free with unlimited dictations, Unlimited Notetaker history, Shared dictionary, snippets and team workspace, Centralized billing, usage analytics and priority support |
| Growth | $23/mo | Everything in Pro with unlimited Notetaker access, Single sign-on with SAML, Org-wide HIPAA enforcement with a signed BAA, Admin control over model training and individual usage reporting |
| Enterprise | Custom, annual terms | Everything in Growth, with model training locked off, User provisioning with SCIM, audit logs and managed deployment, Domain capture, invoicing and custom contract terms, Enterprise-level support |
The plan prices are for monthly billing, and the page's annual toggle takes 20% off the per-user price. Free dictation is capped weekly, and the pricing page gives no number for the cap. Growth members get full Notetaker access through 30 November 2026 at no added cost and then move to a limited plan unless upgraded, and Enterprise is available on annual terms only. Checked against Wispr's own plan page on 17 Nov 2026.
| Feature | Free | Pro | Growth | Enterprise |
|---|---|---|---|---|
| Dictation limit | Weekly word limit | Unlimited | Unlimited | Unlimited |
| Notetaker | Limited access | Unlimited history | Unlimited access | Unlimited access |
| Shared team workspace | No | Yes | Yes | Yes |
| SSO with SAML | No | No | Yes | Yes |
| HIPAA BAA | No | No | Yes | Yes |
| Model training locked off | No | No | Yes | Yes |
| SCIM, audit logs and MDM | No | No | No | Yes |
Wispr Flow is an AI dictation app from Wispr AI that turns speech into cleaned-up text in any app on Mac, Windows, iOS and Android, in 100+ languages. Since September 2026 it runs on Canto, a speech model Wispr built itself. Wispr says its users have written more than 60 billion words on Flow.
Speech cleanup in any app
Turns spoken language into punctuated text without filler words and with grammar fixed, in any field on Mac, Windows, iOS and Android.
Canto speech model
Runs on Wispr's own speech model since September 2026, which Wispr claims beat rival models in its own real-world English test.
Dictionary and snippets
Learns your names and technical terms, and lets teams share a dictionary and snippets on paid plans.
Notetaker for meetings
Records and summarizes meetings, detects who is speaking in Teams gallery view and Google Meet, and answers questions across past calls; Mac and Windows only.
MCP connection to AI tools
Lets Claude, ChatGPT, Gemini or any MCP-compatible tool read meeting transcripts, and Wispr says it does not expose your dictations.
Team and security controls
Growth adds SSO and a HIPAA BAA, and Enterprise adds SCIM, audit logs, managed deployment and domain capture.
Privacy settings
Settings let you turn off model training and cloud storage, and audio is deleted once it is transcribed unless dictation cloud storage is on.
| Tool | Price from | Free plan | Runs on | Rating |
|---|---|---|---|---|
| $15/mo | Yes | iOS · Android · Mac · Windows | 3.9 | |
| $3.75/mo | Trial | Mac · Windows | 3.6 | |
| Free | Yes | iOS | -- | |
| $8.33/mo | Yes | Web · iOS · Android · Mac · Windows | 4.4 |
Proprietary
Wispr Flow Free costs $0. Pro is $15 per user a month and Growth is $23, both 20% cheaper with annual billing, and Growth adds SSO, a HIPAA BAA and model-training controls. Enterprise is custom-priced on annual terms.
Yes. The Free plan covers dictation on all four platforms and limited Notetaker access, and Pro removes the dictation cap. The cap on the free plan is not given as a number on the pricing page.
That is a setting. Wispr's privacy page says turning it off means nothing you say is used to train its models, and on Growth and Enterprise it is off and locked for everyone. Audio is deleted as soon as it is transcribed unless you turn on dictation cloud storage.
Wispr says its security is certified under SOC 2 Type II and ISO 27001, both audited by A-LIGN, and that reports are available through its Trust Center. On Growth and Enterprise an admin signs a Business Associate Agreement in the admin portal to enforce HIPAA settings for every member.
Notetaker detects meetings, takes notes and lets you ask questions across past calls, on Mac and Windows only for now. Connected to Claude, ChatGPT or Gemini over MCP, those tools can read your meeting transcripts, and Wispr says the connection does not give them your dictations.
Wispr Flow is a voice dictation app from Wispr AI. It works in any text field and hands back cleaned-up text: SiliconANGLE's report on its August 2026 funding says it fixes grammar, misspoken words and filler such as "ums" and produces clean prose. Wispr announced a $280 million round led by Menlo Ventures on 17 August 2026, which that report puts at $361 million raised in total.
Flow now runs on Canto, the first speech model Wispr built itself, rolled out to everyone in September 2026. Wispr says Canto had the lowest word error rate in its own test of real-world English recordings against models from Google, OpenAI, AssemblyAI and Deepgram. That is the vendor's own evaluation, not an independent one. Flow supports 100+ languages on every plan, and a personal dictionary helps it with names and jargon.
Wispr Notetaker, a separate product inside the app, takes meeting notes on Mac and Windows, with speaker detection in Microsoft Teams gallery view and Google Meet. It connects to Claude, ChatGPT and Gemini, or any MCP-compatible tool, and Wispr says that connection exposes meeting transcripts and not your dictations. Notetaker is not on iOS or Android yet.
Wispr's privacy page lists SOC 2 Type II and ISO 27001 certification, both by A-LIGN, and encryption in transit and at rest. Transcription always runs in the cloud. Model training is a setting you control, and Growth and Enterprise plans lock it off for the whole organization, with a signed HIPAA Business Associate Agreement on those plans too.
Wispr Flow fits heavy writers who want clean text from speech across several devices. Comparison articles found in a search on 17 November 2026 repeatedly name Superwhisper, Aqua Voice, VoiceInk, Spokenly and Willow as alternatives. On Hokai, Contextli and Google AI Edge Eloquent cover dictation, while Otter.ai and Granola cover meeting notes. See also our best AI voice recorders.
Choose Wispr Flow for cross-device dictation that cleans up your speech and can accept cloud transcription, and an on-device tool when you must work offline. No independent hands-on test was run, and Hokai's ratings rest on Wispr's own pricing, privacy and release-notes pages plus a funding report, all read on 17 November 2026.
Product facts on this page (pricing, plans, integrations, compliance) are read from Wispr AI's own pages and documentation. The verdict is Hokai's reading of them, not the vendor's claim.