Hosted in Switzerland

Your conversations, transcribed in Switzerland.

Helvetia Speech turns live and recorded audio into accurate text, with every speaker identified. Hosted on Swiss servers, and your recordings are never kept.

Get started See how it works
  • Live and recorded audio
  • Speaker identification
  • Pay per minute
Swiss servers
No audio retained
Council meeting
Live
3 speakers detected Example
How it works

From audio to text in three steps

  1. 1

    Create a token

    Sign up and generate an API token from your dashboard.

  2. 2

    Send audio

    Stream a microphone live or upload a recording, with voice samples of your speakers if you like.

  3. 3

    Get text with speakers

    Receive the transcript with timestamps and speaker labels, ready to use in your application.

Features

Everything you need to transcribe conversations

Live transcription

Stream audio over a WebSocket and receive text as people speak. Ideal for meetings, hearings and events.

Recorded audio

Upload a file and get the full transcript in one request, or process long recordings in the background.

Speaker identification

Share a short voice sample per participant and every line is attributed by name. Unknown voices are labelled automatically.

Summaries and drafts

Turn transcripts into summaries, action items or draft minutes with a language model hosted in Switzerland.

Simple integration

REST and WebSocket endpoints, bearer-token authentication and clear JSON responses. No models to host or maintain.

Pay for what you use

Billed per minute of audio and invoiced monthly in CHF. Follow your usage in the dashboard.

Who it's for

Built for people who handle sensitive conversations

Municipalities and administrations

Produce draft minutes of council and committee sessions with every speaker named. Your team reviews and approves, instead of typing for hours.

Healthcare

Dictate notes and document consultations while sensitive patient audio stays on Swiss infrastructure.

Legal

Transcribe hearings, interviews and dictation with speakers kept apart, without sending client data abroad.

Software vendors

Add transcription and speaker identification to your product with a few API calls, and offer your own customers Swiss-hosted processing.

Privacy

Audio goes in. Text comes out. Nothing stays.

Helvetia Speech is built for conversations that cannot leave the country. Audio is processed on Swiss servers and discarded once your transcript is delivered.

  1. Your audio

    Sent over an encrypted connection.

  2. Processed in Switzerland

    Transcription and speaker identification run on Swiss servers.

  3. Text delivered

    You receive the transcript and decide what happens next.

  4. Audio discarded

    Recordings and transcripts are not kept on our servers.

What this means in practice

  • Hosted and processed in Switzerland
  • No retention of recordings or transcripts
  • Only billing data is kept: duration, mode and timestamps
  • Encrypted connections over HTTPS and secure WebSocket
  • Per-token IP restrictions, rate limits and instant revocation
  • Data processing agreement available on request

Helvetia Speech is designed to help you meet your obligations under the Swiss Federal Act on Data Protection (FADP). Compliance also depends on how you use the service.

For developers

A few lines to your first transcript

Authenticate with a bearer token and call an endpoint. The same API powers our own meeting-minutes mobile app.

  • REST for recorded audio, WebSocket for live streams
  • Speaker names from voice samples
  • Timestamps on every segment
  • Usage returned with each response
Get your API token
Response
Pricing

Simple pricing, per minute of audio

No subscription. You pay for the audio you process, invoiced monthly in CHF.

Estimate your monthly cost

Type of audio
20 h
Estimated cost per month CHF 144.00 CHF 0.12 per minute · CHF 7.20 per hour

Indicative estimate based on the rates shown.

Transcription With speaker identification
Recorded files CHF 0.09 per minute CHF 0.12 per minute
Live streaming CHF 0.14 per minute CHF 0.17 per minute

Summaries and drafts

Language model, billed per 1,000 tokens

CHF 0.20

Processing a lot of audio? Ask us about volume pricing.

Talk to us
FAQ

Questions, answered

Are my recordings or transcripts stored?

No. Audio is processed on Swiss servers and discarded once your transcript has been delivered. We only keep the data needed for billing: audio duration, mode and timestamps.

Which languages are supported?

French is our primary language today. If you need other languages, get in touch and tell us about your use case.

How does speaker identification work?

Provide a short voice sample for each participant. Helvetia Speech then labels every segment with the matching name. Without samples, speakers are numbered automatically.

Can I use it for official minutes?

Use it to produce a draft. A person should always review and approve the text before it becomes the official record.

Is Helvetia Speech compliant with the FADP?

The service is designed to help you meet your obligations: Swiss hosting, no retention and a data processing agreement on request. Whether your overall use is compliant depends on your own processing, so we recommend involving your data protection officer.

How do I get started?

Create an account, generate an API token in the dashboard and send your first request. If you'd like help with the integration, just ask.

Hear it on your own recordings

Create an account, generate a token and send your first audio file today.

Create your account Talk to us