Kizwork — the sovereign AI meeting platform that proves who is speaking

Your voice, in every language.And only yours.

Anyone can now clone a voice in a few clicks. Kizwork makes sure the voice on your call is human — and truly who it claims to be.

  • Speech-to-speech in 47 languagesIn your own voice
  • Real-time voice deepfake detection99.1% accuracy, alert in 2–⁠3 s
  • Certified voice identityProof that it is really you speaking
  • 100% EU cloud or on-premisesAir-gapped if needed, CLOUD-Act-free
Try Kizwork free
How exposed are your meetings?Sovereignty risk score: 12 questions, instant result, no sign-up.Take the scorecard
Who it is for

Three sectors where a fake voice comes at a high price

Same platform, different stakes. Pick yours.

Executives meeting in a bright boardroom

Banks & insurers

Payment approvals run over calls. Every voice checked in the meeting, certified identity for the people who can authorise a transfer, Article 50 evidence by default.

The voice that authorises the transfer →

 Operators at their screens in a bright operations room

Gendarmeries, police & defence

Joint operations across seven forces and languages, one deployment per force, air-gap capable. Sources protected by irreversible anonymisation.

Seven forces, seven perimeters →

 Founder smiling on a video call in a sunny office

Growing companies

Secure video, AI minutes, live translation in your own voice and deepfake defence — one tool instead of four, from €7.99 a seat.

Four invoices become one →

And for everyone else: understand each other in 47 languages, in your own voice

Students, freelancers, small teams, estate agents, doctors, lawyers: most of the time, the point is simply to understand each other. Kizwork translates what each person says into the other's language, live, and everyone keeps their own voice. Meetings flow, relationships stay personal, and nobody has to talk through a robotic narrator.

  • Students
  • Freelancers
  • Small teams
  • Real estate
  • Doctors
  • Lawyers
  • Enterprises

Your own voice after a one-time recording of about 3 minutes; a standard voice until then. Included in the free demo.

See speech-to-speech in action →

Incubated company · Fall Cohort 2026

Established in France and Portugal · Web Summit Lisbon 2026 ALPHA

Why Lisbon →
Why Kizwork

One platform for meetings that cannot go wrong

Most meeting platforms were built for convenience. Kizwork was built for control.

Sovereign by design

EU infrastructure (Scaleway — Paris, Amsterdam, Warsaw), outside the reach of the US CLOUD Act, or your own servers. Your data, your jurisdiction, your keys.

Meeting intelligence, built in

Real-time transcription and AI reports — summary, decisions, action items — in 47 languages.

Multilingual without friction

Live transcription and translation tuned to your industry vocabulary, for every participant.

Trust every voice

Every voice in the meeting is checked live, by default. Certified voice identity and voice anonymisation are live today in the Voice Trust+ add-on.

A real-world scenario

The voice on the call sounds like your CFO. Is it really them?

A video call, an urgent transfer, a familiar voice. In 2024 an employee at Arup in Hong Kong sent about US$25 million after a meeting in which the CFO and colleagues were deepfakes.

How Kizwork stops deepfake fraud

What happens in a Kizwork meeting

  1. In 2–⁠3 seconds, the cloned voice is flagged as synthetic on screen.
  2. A voice challenge (2FA) asks the real CFO to confirm their certified voice identity.
  3. The payment waits for proof, not for a familiar voice.
Speech-to-speech

Speak your language. Be heard in theirs — in your own voice.

Every participant hears you in their language, across 47 languages, with your voice, your tone and your emphasis — not a synthetic narrator reading out a translation.

  • Your own voice. Recorded once, in about 3 minutes. Tone and emphasis survive the language barrier.
  • No third-party AI voice. No external voice service in the loop, and no outside provider holding a model of your voice.
  • Marked and consented. Every voice we synthesise is marked as synthetic, with the speaker's explicit consent (AI Act Article 50).
  • Backed by certified voice identity. Listeners know who is really speaking behind the translated voice.

In the Voice Trust+ add-on (+€4.99/user/month on Starter or Pro Security) — and in the free demo.

Try Kizwork freeHow speech-to-speech works
Outcomes

Meetings that end with decisions, in every language

  1. Decisions written up for you

    AI minutes record decisions and owners, so nobody has to take notes.

  2. Nothing lost in translation

    Searchable, multilingual transcripts for every meeting, in every participant's language.

  3. Answers your DPO expects

    EU jurisdiction, GDPR-native, aligned with NIS2 and DORA.

Technology Alliance

Best-in-class European voice AI, one sovereign platform

Award-winning European voice technology, integrated into one platform and hosted in the EU.

Voice AI · French deeptech · ICCA 2026 winner Live transcription and translation tuned to your business vocabulary · deepfake detection, noise reduction and speech enhancement · powers the Verified Voice shield, live in production.
Sovereign European cloud ISO 27001 & HDS-certified data centers · Paris, Amsterdam, Warsaw · CISPE & GAIA-X member · foundation of the Kizwork Sovereign Loop.
The sovereign platform Secure video, AI minutes, multilingual intelligence and voice trust — unified, EU-hosted, governed by European law.

Explore the technology alliance

The jurisdiction problem

The alternative to Teams, Zoom and Meet — without the CLOUD Act

US platforms offer "EU data regions." But the US CLOUD Act still applies to them, wherever your data sits. Sovereignty is not a checkbox — it's an architecture.

CriterionKizworkUS platforms
Data jurisdiction100% EUSubject to US CLOUD Act
AI processingIn-region, self-hosted modelsExternal / US-routed APIs
Key managementCustomer-managed keysProvider-controlled
Voice authenticityDeepfake defence built in, on by defaultNot available
Supply chainEuropean stackUS hyperscalers

Visit the Trust Center

See it live

See Kizwork flag a synthetic voice

A 30-minute private demo, in your language and on your use case.

Try Kizwork free