A man sits at a wooden table in a sunlit loft, speaking into a desk microphone
Voice · spoken, then sorted

Say it once.
Klausi knows where it goes.

Dictate in 36 languages, straight into notes, tasks, and invoices. Klausi recognises client, project, and service from what you said — natively in the Mac menu bar.

Voice in numbers

36dictation languages
EUspeech recognition in the EU
7quick actions for notes
9quick actions for tasks
Previewbefore anything is applied
72 hmeeting audio, then deleted
10 minvoice minutes on the free plan
300 minvoice minutes with Pro
macOS 13+Mac app, notarised
nowake word

Others hand you text back. Klausi knows where it belongs.

Four stations between the spoken sentence and the finished entry — and you confirm every one of them.

  1. 01Dictation

    Speak your language. In 36 of them.

    Notes, tasks, time recaps, the Klausi bar: the same microphone everywhere. Audio goes encrypted and directly to speech recognition in the EU — all that reaches us is the duration, not the wording.

  2. 02Quick actions

    One click sharpens, what you said.

    Seven actions for notes, nine for tasks: optimise, summarise, TL;DR, open questions, in English, as an email. Every result arrives as a suggestion first — you decide whether to apply or discard it.

  3. 03Assignment

    Client, project, service — read from your sentence.

    Klausi matches what you said against your real records: client from your client list, project from your projects, line item from your service catalogue including its stored price. If several fit, Klausi asks instead of guessing.

  4. 04Meetings

    The conversation writes itself down. In the right project.

    The Mac app notices from the microphone status that a conversation is running, and asks once. It records microphone and system audio — no screen. The transcript lands as a note wherever you file it.

An honest comparison

Wispr Flow replaces your keyboard. Klausi replaces the follow-up work.

Both are good tools — they simply solve different problems. One gets words onto the screen. The other gets them where they belong.

Wispr Flow

The better keyboard

  • Types into any app without a plugin — Slack, Gmail, terminal, editor
  • 100+ languages, Mac, Windows, iPhone, and Android
  • Meeting notes with speaker identification, calendar and Slack integrations
  • Afterwards the text sits in the field you spoke into

Klausi

The system behind it

  • 36 languages, straight into a note, task, time entry, or invoice
  • Client, project, and service come from your own real records
  • “Extract tasks” creates real tasks — attached to the project, not a list of text
  • Your spoken sentence becomes an invoice draft with finished line items
  • Speech recognition in the EU, meeting audio deleted after 72 hours

What Klausi can’t do: type into other apps. “Dictate & copy text” puts what you said on your clipboard — pasting is up to you. And we don’t integrate with Slack.

Wispr Flow details as stated on their own website, as of August 2026.

From the beta

What first users appreciate about Klausi.

No big promises. Just clear reactions to a tool that works more calmly.

I finally stopped tracking hours in the Notes app.
Toni S. · Videographer
For the first time, my tax advisor had no questions.
Julia W. · Copywriter
Travel expenses were always the last thing I did. Now it just happens on the side.
Marco D. · Consultant
One tool less on the bill. And still more overview.
Nadia K. · UX designer
With Klausi I start the morning with an automatic plan instead of five open tabs.
Miriam K. · Design consultant
I no longer piece client info together from three tools and two lists — and I know the next step.
Leonie R. · Brand consultant

On the Mac

Native in the menu bar. Not in a tab.

The Mac app is signed and notarised by Apple and runs on Apple silicon from macOS 13. It lives in the menu bar, and the capture bar floats above every window and desktop — without stealing your focus.

Included in every plan, the free one too.

Feature overview

Everything voice can do in Klausi. Honest, no fine print.

Dictation

  • 36 languages, switchable anytime
  • Your own shortlist of the languages you actually use
  • Live text while you are still speaking
  • A short pre-roll so your first words aren’t lost
  • Pick the microphone straight from the menu bar
  • A hard stop per place: 5, 2, or 1 minute

Where you can speak

  • Notes — dictation and quick actions in the editor
  • Tasks — voice task with no intermediate dialog
  • Time tracking — recap after the timer stops
  • Home — speak your question to the AI assistant
  • Dictation to copy on the Mac — text to the clipboard

Quick actions

  • Notes: optimise, summarise, extract tasks, TL;DR
  • Notes: open questions, in English, as an email
  • Tasks: optimise, checklist, steps, estimate effort
  • Tasks: definition of done, TL;DR, open questions
  • Improve, shorten, correct, or clarify selected text
  • Preview with discard and apply before every change

Assignment

  • Extracted tasks inherit the note’s project
  • Time recaps inherit project and client from the entry
  • Client recognition for invoices and proposals
  • Service recognition including the stored catalogue price
  • Project recognition — the client follows the project
  • Ambiguous? Klausi asks instead of guessing
  • Unknown name? One click creates the client

Meetings on the Mac

  • A prompt as soon as another app uses the microphone
  • Microphone and system audio — no screen
  • Title suggestion from a running booking or the detected app
  • File it privately, with a client, or in a project
  • Transcript and note stay; audio goes after 72 hours
  • Free 60, Pro 600 meeting minutes per month

Trust & integrations

  • No wake word — recording only starts when you do
  • Speech recognition in the EU, one-time keys instead of standing access
  • Training opt-out set with every speech provider
  • Microphone permission only at your first recording
  • Use your notes in Claude, ChatGPT & co. (MCP)
  • Free 10, Pro 300 voice minutes per month

Belongs together

Speaking is only the start.

What you say flows on into notes, projects, time, and invoices.

Frequent questions

Questions about voice in Klausi.

Who is this for?

For the self-employed and small teams who talk all day — with clients, in the studio, on the road — and still have to document it in the evening. If your head is faster than your hands, you’re in the right place.

Which languages does Klausi understand?

Thirty-six, from German, English, Turkish, and Spanish through to Japanese, Ukrainian, and Swiss German. You pick the language per recording and keep a shortlist of the ones you actually need — the rest stays out of your way.

Can I dictate into other apps with Klausi?

Not directly — and we say so openly. Klausi does not type into other windows. On the Mac there is “Dictate & copy text” for exactly this: Klausi records, you get the finished text on your clipboard — pasting is up to you. That dictation is not stored in Klausi. Inside Klausi, on the other hand, what you said lands exactly where it belongs: in the note, the task, the time entry, or the invoice draft.

How does Klausi recognise client and project?

Klausi matches what you said against your own records — your client list, your projects, your service catalogue. If exactly one entry fits, it gets set; if several fit, you get them as options instead of a guess. Name a client that doesn’t exist yet and one click creates it. Nothing is written until you confirm.

Is Klausi listening?

No. There is no wake word and no background service recording audio — you start every recording yourself. On the Mac, Klausi checks every twelve seconds whether another app is using your microphone, in order to offer you a meeting recording. No audio is captured in the process, only an operating-system device status is read — and the orange dot in the menu bar shows you any microphone use anyway.

What happens to what I say?

When you dictate, your device opens a direct, encrypted connection to speech recognition in the EU — the audio passes through no Klausi server, and all that reaches us is the duration. Meeting recordings are different, because we process the audio for you: it sits in our EU storage for at most 72 hours and is then deleted automatically; the deadline is enforced in the database. Training opt-out is set with every speech provider — you don’t have to go looking for it.

What does it cost?

Getting started is free: ten voice minutes and sixty meeting minutes per month. With Pro it’s three hundred voice minutes and six hundred meeting minutes. The Mac app itself costs nothing extra — it belongs to every plan, the free one included.

Do I need a Mac?

Not for dictation — notes, tasks, time recaps, and asking the assistant all work in the browser too. Two things are Mac-only: the menu bar with its floating capture bar, and meeting recording. There is no Windows version.

Your next thought
doesn’t belong nowhere.

Download the Mac app and get going — dictation, quick actions, and meeting notes are included right away.

Free · 36 languages · Speech recognition in the EU · no wake word