Private beta · EU-hosted

The world speaks
your language.

A private browser-based voice bridge that translates your friends in real time. Speak naturally, hear each other in your own language — powered by OpenAI GPT-Realtime-Translate. No downloads, no recordings.

AI translation can make mistakes. Not for emergency, medical, legal, or professional interpretation.

BerlinDeutsch
LisbonPortuguês
ParisFrançais
One conversation · three languages

What LinguaCall does

A voice bridge built for the way friends actually talk.

Personal voice for every listener

Each participant hears others in their own language through a clearly synthetic voice. Same conversation, three languages, no interruption.

Invite-only rooms

No accounts. Share a secure link with two to six friends. The room disappears when the call ends — no recording, no transcript.

EU-hosted, GDPR-first

All processing happens in approved EU regions. Audio is ephemeral. Content is never used to train models.

How it works

Three quiet steps between hello and understood.

01

Create a room

Choose your spoken and listening languages, test your mic, and share a secure invite link with your friends.

02

Speak naturally

Talk in your own language at your own pace. Overlaps, laughter, and small pauses are handled gracefully.

03

Hear each other

Every listener gets their own translated audio stream, synced with the active speaker in the room.

Privacy by design

Private conversations, protected in Europe.

No recordings

Audio, transcripts, and translations are ephemeral by design.

EU data residency

Processing stays in approved EU regions with documented subprocessors.

Consent-first

Explicit AI-processing consent before your microphone is ever transmitted.

No training on your calls

Your conversations are never used to train general-purpose models.

Supported today

Understand 70+ languages. Speak 13.

Powered by OpenAI GPT-Realtime-Translate: automatic detection of 70+ source languages, translated into any of these 13 listening languages.

EnglishDeutschFrançaisEspañolPortuguêsItalianoРусскийहिन्दीBahasa Indonesia日本語한국어中文Tiếng Việt

Questions

Straight answers.

Do you record calls or store transcripts?+

No. Audio, translation buffers, and any temporary text representations are cleared as soon as they are no longer needed to deliver the call.

Does the translated voice sound like me?+

No — and that's intentional. LinguaCall uses stable but clearly synthetic voices per speaker. It never clones or imitates your biometric voice.

How many people can join a call?+

Between two and six participants per room. Each participant picks their own spoken and listening languages independently.

What happens if translation lags or fails?+

You'll see a clear status message and, depending on your fallback setting, hear the original audio at reduced volume until translation resumes.

Can I use this for emergencies or legal interpretation?+

No. LinguaCall is built for informal private conversations, not safety-critical, medical, legal, or professional interpretation.

Built for private conversations, not data harvesting.

"My mother in Warsaw and my partner in Lisbon finally had a real conversation. No apps, no dictionaries — just each other."

— Anna K., early tester

Ready to break
the silence?

Start a call today. No registration, no downloads — just a secure link for you and your friends.

Start a call