A dictation app is reportedly raising $260M at a $2B valuation. Wispr Flow, in case you missed it. Six months ago they were at $700M; this round (led by Menlo Ventures) nearly triples that.
If your first reaction is "wait, why is dictation worth this much?" — fair. OpenAI Whisper has been free for two years, SuperWhisper is already a familiar local app. So what's different?
So why is it $2B?
Cut to the answer: Wispr isn't selling "dictation," it's selling "transcribe-and-clean-up". Looks like the same voice input on the surface, but the output is a different product.
Old-school dictation: you say "uh, so, um, can we move the meeting to next Tuesday" and the screen shows exactly that. You then go in and delete the filler, fix the sentence, format it. The transcription was fast, but the editing was its own job.
Wispr Flow takes the next step. The same sentence lands on screen as "Can we move the meeting to next Tuesday?". Filler words removed, sentence cleaned, learned terminology applied, tone adjusted to whatever app you're in (email, Slack, code editor) — automatically.
Notable Capital (an investor) flagged a metric worth chewing on: users hit Enter 0.5 seconds after the dictation lands. That means the "re-read, fix it up" step disappeared. Trust that you don't need to touch the output, measured in half a second.
Why now is the inflection point
Whisper has been open source since 2022. Accuracy was already good. The dictation market actually went vertical in 2025-2026 — once LLMs were cheap and fast enough to clean up the transcript on the fly. STT got solved. The category quietly moved from "speech to text" to "speech to a finished sentence."
The $2B in numbers
Bloomberg broke the round — terms aren't final yet — but the interesting datapoint isn't the dollar number. It's that 60% of usage is non-English. India is the #2 market behind the US (14% of installs), and Hindi, Mandarin, and Spanish are growing fast. The harder your language is to type on a Western keyboard, the more value "just talk" creates.
Customer base tells a story too. Nvidia and Amazon engineers reportedly use Wispr Flow to drive coding assistants and workplace tools. There are dedicated extensions for Cursor and Windsurf. SOC 2 Type II and HIPAA compliance shipped across all plans, which is the lubricant for enterprise adoption.
What's the real difference vs Whisper and SuperWhisper?
Accuracy is roughly a wash. The split is "who cleans up the transcript after".
| Whisper / SuperWhisper | Wispr Flow | |
|---|---|---|
| Output | Raw transcription ("uh… um…") | Filler-stripped, polished sentence |
| Where it runs | Local (offline OK) | Cloud (needs connection) |
| Platforms | Mac-centric (Whisper does Linux too) | Mac · Windows · iOS · Android |
| Monthly cost | $0–$9 | $15–$18 (Pro) |
| Enterprise | Personal use mostly | SOC 2 · HIPAA out of the box |
SuperWhisper has its own clear lane. Local processing means it works offline and the audio never leaves your machine, which is real for privacy-sensitive work (legal, parts of healthcare, internal-only material). Half the monthly price too. The catch: Mac-only.
Whisper itself is even simpler. Open source, a $29 one-time desktop app (get-whisper) exists, accuracy is effectively at parity. If you only need "transcribe, that's it", there's no reason to pay for a cloud subscription.
Wispr Flow doesn't sit between them — it sits one level up. It's not selling "a dictation tool," it's selling "finish a text input via voice". That's why the output drops into email, Slack, CRMs without needing a second pass.
What non-English speakers should know
Wispr supports 104+ languages, but eight of them (English plus German, French, Spanish, Portuguese, Italian, Hindi, Thai) are at "model parity" — the rest, including most Asian languages outside Hindi, are a tier below. Short sentences land well; jargon-heavy or grammar-complex inputs may still need editing. SuperWhisper has similar behavior. If your workflow is heavily non-English, test on real sentences during the free trial before committing.
How to actually start
- Define your workflow in one sentence first
"70% of my typing is email and Slack replies" vs "I dictate long document drafts" vs "I want voice commands in my code editor" lead to different tools. - Decide local vs cloud upfront
If privacy and offline matter, SuperWhisper or Whisper (get-whisper). If auto-editing, multilingual, or enterprise compliance matter, Wispr Flow. - Use the 14-day trial as a real test
Wispr Flow gives 14 days of Pro free, no card needed. Pick 10 sentences you actually use in your job and measure how often you have to touch up the output. - Set up hotkeys and Snippets first
Wispr's Snippet Library turns frequent phrases (signatures, meeting URLs, stock intros) into voice triggers. 30 minutes of setup changes how the tool feels. - Pull it into your code editor
Cursor and Windsurf have dedicated extensions. The highest-ROI pattern is dictating long prompts into the AI chat instead of typing them — users consistently report 3–4x faster than the keyboard.
Resources to go deeper
The Bloomberg scoop The original report on the Menlo-led $260M / $2B round bloomberg.com
Wispr Flow official site Demo, pricing, IDE extensions, and the 14-day Pro trial in one place wisprflow.ai
Notable Capital deep dive The "0.5-second trust" metric, the BCI-to-software pivot, and the 150-year-old keyboard interface argument notablecap.com
Whisper vs SuperWhisper vs Wispr Flow Pricing, platforms, accuracy, offline — a 2026 head-to-head get-whisper.com
Tanay Kothari profile The Stanford-trained founder betting on a voice-first future businesstoday.in
Wispr Flow vs SuperWhisper (Wispr's own comparison) Wispr's own breakdown — SOC 2 and HIPAA across plans wisprflow.ai
