Souffleur

The notes write themselves. The promises get kept.

Souffleur sits quietly in your ear during a meeting. It writes the minutes, pulls out what you promised, and then goes and does it — the follow-up email, the calendar entry, the task in your own systems. Not another recording you will never listen back to.

A prompter, not a participant

In the theatre, the souffleur feeds you the line you forgot without ever stepping on stage. That is exactly the job. It does not join your meeting, interrupt it or speak for you. It listens, and afterwards the work is already done.

During

It listens

At the table it uses your phone. In an online meeting it also takes the audio from your computer, because the other person's voice is in the machine, not in the room. On a call, it works through the phone system itself.

Ask

It answers, quietly

Forgot what was agreed twenty minutes ago, or what this client ordered last year? Ask, and the answer arrives in your ear. Nobody else at the table hears it.

After

It acts

The summary is the easy part. Souffleur separates what was merely discussed from what you actually promised, and then carries it out in the systems you already use.

Anyone can transcribe. Almost nobody follows through.

There is no shortage of tools that record a meeting and hand you a summary. The summary is not the problem. The problem is the sentence "I will send that over tomorrow" — said in passing, forgotten by the afternoon, remembered by the customer.

Souffleur is built around that sentence. It is connected to real systems, not to a notes app: your calendar, your contacts, your invoicing, your mail. A commitment becomes a task with a date, and it comes back to you until it is done.

Built for European rules, not around them

Recording a conversation is not a neutral act, and Europe is not one jurisdiction. In Germany, recording without the consent of everyone present is a criminal offence. Tools that leave that entirely to the user are shifting a legal risk onto their customer.

Per meeting, never always on

There is no ambient mode that listens to your day. You start it for a specific conversation and you stop it. That is a deliberate limit, not a missing feature.

Consent belongs in the product

Announcing that a meeting is being assisted should take one tap, not a footnote in a policy document. Disclosure is designed in, because for business use it has to be.

Your recordings, your systems

Transcripts land in your own environment. There is no shared corpus, no training on your conversations, and no third-party portal you have to keep paying to read your own history.

You make the hardware. We make it worth wearing.

Smart glasses, earbuds, badges, pendants, meeting speakerphones — the hardware coming out of Asia is good, and it is getting cheaper every quarter. What decides whether it sells in Europe is rarely the hardware. It is the software on top of it, and whether the product is allowed on the European market at all.

What we bring

The assistant layer

A working service, not a slide deck: capture, transcription, summarisation, commitment extraction and follow-through. Your device streams audio and gets structured results back.

How it ships

White label or embedded

Under your brand, in your app, with your look. Or embedded through an interface, if you already have an app and only need the intelligence behind it.

Why Europe is different

Market access, handled

A connected product sold in the EU falls under the Cyber Resilience Act, with reporting duties already in force. We run a compliance platform for exactly that. Software and market access from one partner.

What a partnership actually looks like

  • You keep the customer. Your brand, your channel, your pricing. We are the layer underneath.
  • We keep the software honest. European hosting, European data rules, consent designed into the flow.
  • One integration, three markets. Phone, laptop and wearable already share the same service, so a new device is a new client, not a new product.
  • Start small. A pilot on a single device model, with a working demo in weeks rather than quarters.
We already have an app. Do we have to replace it?
No. The intelligence sits behind an interface. Your app keeps its own identity and simply talks to it.
Does the audio have to leave the device?
Audio is processed to produce text and then it is your data, in your environment. What is kept, and for how long, is a setting — not something you discover afterwards.
What does the Cyber Resilience Act mean for us?
If your product connects to anything and is sold in the EU, it applies to you, and the reporting obligations are already live. Get in touch and we will tell you plainly what it means for your specific product.

Tell us what you are building

Whether you make a device and need the software, or you want to try the assistant yourself first — write to us and you will get a real answer from a person, not a form response.

info@souffleur.online