Wispr Flow pricing in 2026: the free tier, the real annual cost, and the $0 alternative

Wispr Flow costs $0 on the Free plan, $15 per user per month on Pro billed monthly, or $12 per user per month on Pro billed annually. Enterprise is quoted on request and shows no figure. I read all of that off Wispr Flow's own pricing page on 19 August 2026. The number most pages leave out is the total: three years of Pro comes to $540 at the monthly rate or $432 at the annual rate, and the free plan stops at 2,000 words a week on desktop.

Pricing pages are the one place a vendor cannot be vague, so they are worth reading directly instead of through a listicle. What follows is every figure Wispr Flow publishes, what each plan restricts, and the multiplication nobody puts in a headline.

How much does Wispr Flow cost in 2026?

Three plans, two of them priced. Free is $0 a month. Pro is $15 per user per month billed monthly, or $12 per user per month when you switch the page to annual billing, a toggle labelled "20% discount". Enterprise says "Contact us" and carries no figure anywhere on the page. All of that was read on 19 August 2026.

Pro is sold to individuals and teams from the same tier, and its feature list is short: unlimited dictations, more meetings kept for longer, advanced AI thinking models, early access to new models, priority support tickets, and centralised billing. Everything else sits in the free column, including speech to text in more than 100 languages, the custom dictionary, and the Notetaker.

Two pages ranking for this query today describe a tier Wispr Flow does not currently show. Voibe's pricing explainer lists a separate Teams plan with a three-seat minimum, and Zack Proser's pricing guide describes a 14-day trial. Neither appears on Wispr Flow's own page as I read it, and its FAQ says a Pro team can be created "with no minimum seat requirement".

What does the Wispr Flow free tier actually limit?

Dictated words, and the ceiling changes by device. The plan comparison lists 2,000 words a week on desktop, 1,000 a week on iPhone, and unlimited on Android. Meetings carry a weekly limit the page never quantifies, and note retention is described only as "Standard" against Pro's "Extended". Features are barely restricted at all.

The desktop cap is the one that decides whether you upgrade. Spread over five working days, 2,000 words a week is roughly 400 words a day, which is one long email plus a couple of Slack replies. Over a year that comes to about 104,000 words, their weekly figure multiplied by 52. Anyone who dictates as their main input method will pass it early in the week.

What the free plan does not restrict is worth stating plainly, because it is unusually generous: more than 100 languages, the personal dictionary that learns names and jargon, the Notetaker on Mac with speaker identification, MCP access so notes work in Claude and ChatGPT, calendar and Slack connections, and opting out of model training at any time.

Where does Wispr Flow genuinely win?

On hardware that cannot run a speech model, and on finish. Because transcription happens on Wispr Flow's servers, an ageing ultrabook or a low-power office machine gets the same result as a new workstation, with nothing to download and no GPU to own. It also ships iPhone and Android apps, and a support queue with a person behind it.

That is a real advantage and it is the honest reason to pay. A local model has to fit in your RAM and run on your silicon; a cloud model does not care what you own. Wispr Flow lists dictation on Mac, Windows, iOS, and Android, with the Notetaker on Mac only for now. Its iPhone app holds a 4.8 average from about 14,000 ratings on the US App Store listing, where it sat at number 65 in Productivity when I checked.

There is also a difference in resources no comparison table shows. Wispr Flow's about page describes "a team of designers, AI researchers, and engineers" and names backers including Kevin Weil, chief product officer at OpenAI. SpeakoFlow is one person. That shows up in polish, in mobile coverage, and in how fast bugs get fixed.

"I've never seen a tool adopted this fast. In particular across both sales and eng." Tuhin Srivastava, co-founder and chief executive of Baseten, quoted on Wispr Flow's business page.

What does Wispr Flow cost over one year and three years?

Wispr Flow does not publish an annual total, so this is arithmetic on its own monthly figures. At the annual rate of $12 per user per month, one user costs $144 a year and $432 over three years. At the monthly rate of $15, the same user costs $180 a year and $540 over three years.

The chain, so you can check it: the pricing page states $12 per user per month on annual billing, and 12 times $12 is $144, times three years is $432. It states $15 per user per month billed monthly, and 12 times $15 is $180, times three is $540. Annual billing therefore saves $36 a year, and a five-person team on it works out at $720 a year. None of those totals appear on Wispr Flow's page; they are multiplications of the prices that do.

Is there a student discount, a team plan, or a business price?

Yes to the first, partly to the second, and no to the third. Wispr Flow's students page lists Flow Pro for Students at $6 a month billed annually, described as 50% off, with three months free for new subscribers. Teams run on the ordinary Pro tier. Enterprise pricing is quoted by sales only.

Verification is automatic with an education email address, and the students page says a student ID, enrolment confirmation, fee receipt, or dashboard screenshot works instead. The pricing FAQ mentions "discounted Flow Pro to nonprofit organizations" without stating the discount, so that figure is not published.

For teams, the split matters more than the price. Pro covers shared dictionary and snippets, centralised billing, and basic usage dashboards, with no minimum seat count. Wispr Flow's business page puts SOC 2 Type II compliance on Enterprise plans only, lists ISO 27001 as something the company is working toward rather than holding, and offers HIPAA readiness on all plans once a business associate agreement is signed. If your procurement checklist has SOC 2 on it, your price is a conversation rather than a number.

What is the subscription actually paying for?

Server-side inference, mobile apps, a meeting notetaker, and a support queue. Every dictation you speak costs Wispr Flow compute on its own hardware, which is the structural reason a cloud dictation tool needs a recurring fee at all. Pro adds unlimited volume, longer note retention, advanced models, and priority support.

The trade you accept is where your audio goes. Wispr Flow's transcription runs on its servers by design, which is exactly what makes it work on hardware that could never host a model. Its pricing page promises that your data stays private and HIPAA ready, and that you can opt out of model training at any time. Apple's App Privacy panel on the iPhone listing is the more sobering read: it records that identifiers and usage data may be used to track you across apps and websites owned by other companies, and that contact info, identifiers, usage data, and diagnostics may be linked to your identity.

Neither of those is a measurement I made, and neither is an accusation. They are two published positions from two primary sources, both worth reading before you decide what $432 over three years is buying.

What does the same dictation job cost at $0?

Nothing, if your machine can run a 600 million parameter speech model. SpeakoFlow is MIT licensed, needs no account, has no word cap, and transcribes on your own device on Windows, macOS, and Linux. It trains no models of its own, so its accuracy is whatever the published open models score.

The default is NVIDIA Parakeet TDT 0.6B v3, which records a 6.34% average word error rate and 1.93% on LibriSpeech test-clean on the Open ASR Leaderboard. On multilingual short-form audio, the leaderboard paper (arXiv:2510.06961) puts that same model at 4.81% average WER and 1720 RTFx against OpenAI Whisper Large v3 at 4.81% and 111 RTFx: identical measured accuracy at roughly 15 times the throughput. On long-form English the same paper has Parakeet at 10.7% WER, ahead of Whisper Large v3 at 11.2%.

Those throughput figures come from datacenter GPUs with batching, so a laptop is far slower in absolute terms and I do not claim otherwise. The ratios hold, which is why finished text lands as soon as you stop speaking. The real costs of the $0 route are different in kind: a multi-gigabyte model download, hardware that can run it, no mobile app, and a macOS build that needs one Terminal command on first launch because it is not yet signed by Apple.

For the feature-by-feature version rather than the money, our SpeakoFlow and Wispr Flow comparison covers privacy, offline behaviour, and platform support. For the wider field, including the Mac-only open-source tools, see the best free and open-source Wispr Flow alternatives.

Frequently asked questions