Dictation and transcription in Switzerland: reports and minutes without the audio going abroad
Dictation is back: speak the report on the way out of a meeting, and it's text by the time you're at your desk. The app that does it sends the recording to a server you've never seen, to be transcribed by a company you've never met. Your client's name is in the first sentence.
The short answer: speech recognition models such as Whisper run on servers in Switzerland, Meow's or your own, so dictations and recordings are transcribed without the audio going to a foreign provider. A person still reads the text before it counts, because names, figures and technical terms are where transcription slips. How well it copes with your voices and your vocabulary is something to test, on your recordings.
Checked against the law on 30 September 2026
Where does the audio go?
- Audio stays
Recordings are transcribed on servers in Switzerland, operated for you, and can be deleted as soon as the text is checked.
- Your vocabulary
Names, places and technical terms can be given to the model, so it spells them the way your firm does.
- Consent first
Recording a non-public conversation you take part in needs the others' consent; doing it without is a criminal offence (art. 179ter SCC), with narrow exceptions such as business phone calls about orders (art. 179quinquies). Dictating your own notes is not a conversation.
How to test it
0 of 8 checked
Dictations, voice notes, a meeting where everyone agreed: with the accents, noise and speed you really have.
Type out a few minutes of each by hand, or take text you already have, to compare against.
Names of clients, places and products, technical terms, abbreviations.
Transcribe the recordings and count the errors that change the meaning apart from spelling slips.
French, German, Swiss German, Italian, English, and switching between them in one recording.
Note how long a ten-minute recording takes, and decide whether that suits the work: while someone waits, or overnight.
Delete it once the text is checked, unless you have a reason to keep it.
For meetings and calls, the consent rules apply: the guide on AI meeting notes covers them.
Ticked them all? The kit turns this into a test you can run next week: what to check by kind of recording, a word list to start from, the test plan, and what to count before sizing a server. Get the kit
What it handles, and what to check
The kit sorts six kinds of recordings. Two of them:
- A dictated report in a quiet room
- Handles it well; read it through for names and figures.
- A voice note recorded outside
- Usable; noise and wind leave gaps to check.
- Swiss German dialect
- In the kit
- A meeting with several speakers
- In the kit
- Technical or legal vocabulary
- In the kit
- A recorded phone call
- In the kit
Why it runs in Switzerland
A recording holds more than the words: voices, background, whoever else is in the room. Transcribed on servers in Switzerland, operated for you, it doesn't go to a foreign provider, and it's deleted when you decide.
What it needs
Speech recognition runs on a server graphics card; how many hours of audio come in per day decides how big. The Meow Academy's planner shows how models fit on hardware.
Where does your firm stand?
The self-check takes two minutes and says which first step fits. Or talk it through in a free 15-minute call.
Book a free 15‑minute call
Pick a time that suits you and confirm it from your inbox. If the time no longer works on our side, you hear from us the same working day.
What happens in 15 minutes
- 0–5 min
Your situation
What AI should do for you, and what it must never see.
- 5–10 min
What's possible
Which tasks can run privately, roughly what it costs, and what should stay human.
- 10–15 min
The next step
A test on your own documents, an assessment, or nothing yet.
Prefer to write? Send a message
Tell us what you'd like AI to do, and what it must never see. We'll tell you honestly what's possible, what it costs and what should stay human. No details yet? A hello is enough.