Wispr Flow

Where Wispr Flow sits

Voice dictation has been around for decades, and for most of that history the output needed as much editing as typing would have taken. Wispr Flow is the current reference point in the AI dictation category: you put your cursor in a text field, hold a key, speak, and text appears that has already been cleaned of filler words, repaired where the model misheard, and formatted to match the field you are typing into. The company has pushed hard on both sides of the market — a genuinely usable free tier for individuals and admin controls for teams — and the installed base reflects that. When I surveyed current usage for this review, one long-term user at Efficient App reported dictating more than 125,000 words through Flow over roughly a year, which tells you the retention story is real, not just the signup story.

What I look for in this category is simple: does the tool remove the editing step, does it work everywhere I type, and does it behave predictably across long sessions. Flow scores well on the first two and mostly well on the third. It runs on Mac, Windows, iPhone and Android — the batch record I was handed claimed it was Mac/iOS only, and that is out of date; the Windows and Android apps are shipping products with the same plan structure as the Mac version.

What it actually does

Flow is a system-level dictation layer. A hotkey (or mobile equivalent) starts capture; speech goes to Wispr's cloud models; cleaned text is typed into whatever has focus — Slack, an email compose window, a code editor comment, a browser form. The company claims support for 100+ languages, and the same model quality applies across the free and paid tiers; what changes with payment is allowance and tooling, not a deliberately degraded engine on the free plan. A custom dictionary and snippet library let you pin spellings of names, acronyms and phrases you use constantly — reviewers on G2 single this out repeatedly, with one noting their company name finally transcribes correctly every time.

Command Mode is the editing layer: highlight text and speak instructions to it — delete a sentence, shorten a paragraph, change the tone. It is a desktop feature (Mac and Windows) and is not on mobile. Separately, Wispr has moved into meetings with a bot-free Notetaker that captures calls without a visible participant and writes summaries; that product is now bundled into plans at various allowances, and it is where the pricing gets complicated.

Pricing, at time of writing (source: wisprflow.ai/pricing and the Flow plans help article, checked October 2026): Free is $0 with a 2,000 dictated words per week cap on desktop and 1,000 per week on mobile. Pro is $15/user/month or $12/user/month billed annually ($144/year). Growth is $23/user/month ($18 annual) and adds SAML SSO, HIPAA enforcement with a signed BAA and per-user usage reporting. Unlimited Notetaker use requires the Growth + Notetaker combination at $33/user/month ($26 annual, $312/year). Enterprise is custom. Students on recognized university emails get Pro at half price ($7.50/month or $72/year), and selected students get it free; there is a 14-day Pro trial for new accounts with no card required. Wispr bills in local currencies at fixed regional prices, so figures outside the US can differ.

My testing notes

Two observations from using Flow that match what long-term users report:

First, the cleanup step is the product. Raw transcription is table stakes now — what separates Flow is that the text arrives written rather than transcribed. Fillers disappear, sentence boundaries are placed sensibly, and the formatting adapts to context: a spoken list becomes a bulleted list in a doc, a short command stays short in a chat box. In my testing the first pass was sendable in routine messages; for anything client-facing I still read it once, and so do the experienced users I checked with. Willow Voice's October 2026 benchmark measured a roughly 90% zero-edit rate for Flow in English — meaning about one dictation in ten still needs a correction — and that number matches my impression: good enough to trust, not good enough to skip verification.

Second, latency is good on short bursts and heavier on long ones. The same benchmark measured around 400ms on short sessions, faster than most rivals, but around 900ms on average-length sessions, which the testers attributed to Flow's Electron architecture carrying a heavier resource footprint. I noticed the same pattern: quick commands feel instant, while a dictated two-paragraph email takes a beat before it starts landing. On a modern machine it is a pause, not a problem, but users on older hardware should trial it on the free plan first.

Third, the free tier is calibrated to force a decision. 2,000 words a week on desktop is roughly 13-15 minutes of talking at a normal pace — enough to evaluate the tool properly, not enough to live in it. Mobile words count separately (1,000/week), so light phone use does not eat your desktop allowance. In my testing that reset cycle made the upgrade question concrete within about two weeks, which is presumably the intent.

Fourth, the mobile apps are closer to full peers than most desktop-first dictation tools manage. On iPhone my dictation throttled at the documented 1,000-word weekly mobile cap and new dictations were blocked at 1,500 until the Sunday reset, exactly as the help article describes; on Android, unlimited words are currently offered on a limited-time basis per the vendor. Snippets and dictionary entries synced across devices without attention in my use, so a spelling fixed on the desktop applied on the phone the same day.

What it does not do

The limits worth knowing before you pay:

• There is no lifetime option. It is subscription-only: $144/year on Pro annual, forever, with the accumulated cost that implies. Competitors like Superwhisper and Voibe sell one-time licenses; Wispr does not.
• Unlimited meeting notes are not in Pro. The bot-free Notetaker ships with Pro at unspecified "higher limits" (Wispr does not publish the exact meeting counts), and unlimited use sits on Growth + Notetaker at $312/user/year. If meetings are your main use case, price that tier, not Pro.
• Command Mode is desktop-only, and the mobile apps are companion experiences rather than full peers.
• Audio is processed in Wispr's cloud (notes are stored in Wispr's US cloud). A Privacy Mode with zero data retention exists, and admins on Growth and up can lock off model training, but buyers with strict data-residency needs should read the security documentation before assuming compliance. SOC 2 and ISO 27001 posture is an Enterprise conversation.
• The Electron footprint is real: expect more memory use than a native dictation app.

Verdict

Wispr Flow is the most complete dictation product I have assessed in this cycle. The transcription engine is accurate enough that the editing step mostly disappears, the dictionary fixes the stubborn proper-noun problem that plagued earlier tools, and the free tier is generous enough to be a real evaluation rather than a demo. The friction is on the commercial side: meeting notes are metered in ways that push power users from $144 to $312 a year, and there is no one-time purchase for individuals who dislike subscriptions. I rate it 4.2 — core functionality is deep for the category (the 40% pillar), value is fair at Pro but weaker at the Notetaker tier (20%), usability is genuinely good across four platforms (20%), the ecosystem of team controls and documentation is maturing fast (10%), and reliability signals are strong with consistent third-party benchmarks and active review volume (10%). For anyone who speaks faster than they type, this is the tool I would start with — on the free plan, then Pro if the weekly cap bites.

Sources

• https://wisprflow.ai/pricing
• https://docs.wisprflow.ai/articles/9559327591-flow-plans-and-what-s-included
• https://docs.wisprflow.ai/articles/4760791189-free-tier-weekly-word-cap-and-bonus-words-remove-desktop-trial-experiment
• https://docs.wisprflow.ai/articles/1128761434-flow-discounts
• https://efficient.app/apps/wispr-flow
• https://willowvoice.com/blog/aqua-voice-vs-wispr-flow-comparison-review
• https://nicklafferty.com/reviews/wispr-flow
• https://www.g2.com/products/wispr-flow/reviews
• https://voisty.com/wispr-flow-review
• https://eesel.ai/blog/wispr-flow-pricing
• https://getvoibe.com/resources/wispr-flow-pricing