AroerlinkREAL-TIME SPEECH TRANSLATION · LOCAL-FIRST
Speak your language.
They hear theirs.
Aroerlink is a desktop overlay that translates your voice live, inside the meeting apps you already use. It runs on your own device and is built to keep your voice yours.
THE VISION
Language should not decide who is heard.
Speak naturally in your own language and be understood live. Private, because it runs on your own device. Simple, because it works inside the apps you already use.
- EDUCATION
Lectures, classes and campus life where students and teachers do not share a first language.
- COLLABORATION
Teams that work across borders, speaking naturally in the meeting apps they already use.
- PRODUCTIVITY
Fewer follow-ups to clear up what was meant, and more of the meeting spent on the work.
Every day, good ideas are lost in meetings because the person who has them is speaking their second or third language, or is not speaking up at all. Subtitles help you read. They do not let you be heard.
Aroerlink is our answer: real-time speech-to-speech translation that lives on your computer, works with any app that lets you choose a microphone, and keeps your meeting audio on your device by default.
We are building fast, and with the utmost care: every stage measured, every use of AI disclosed, and local-first before anything else. Speed is how we get real value to the people who use it sooner. Care is how we keep their trust.
HOW IT WORKS
One pipeline. Four stages. No cloud in the middle.
You talk. Aroerlink listens, writes down what you said, translates it, and speaks it back as a virtual microphone that your meeting app already knows how to use.
- 01Listen
- 02Transcribe
- 03Translate
- 04Speak
PRINCIPLES
What we will not trade away.
- 01
Local by default
Meeting audio is processed on your own device. Anything that would leave it needs your explicit consent, and the default is that nothing does.
- 02
A voice that stays yours
We are building speech that keeps the speaker's own voice instead of a generic one. It is early work today, and we will say so until it is not.
- 03
Honest about AI
Translated speech is disclosed as synthetic. The overlay says so on screen, in line with the transparency duties of the EU AI Act (Article 50) we are designing toward.
- 04
Fast by design
Latency is a product feature. Every stage is timed and tuned, and nothing is allowed to get in the way of live speech.
ROADMAP
Windows first. The rest on the same engine.
- IN BUILD AND TESTING
Windows
The first platform, with a desktop overlay and a virtual microphone.
- PLANNED
macOS
Follows Windows, on the same engine and the same language packs.
- PLANNED
Linux
After macOS, built from the same codebase and released on the same version.
Aroerlink is in development. Features and timing may change. We do not claim any certification today.
EARLY ACCESS
Be in the room early.
Join the waitlist for early builds and first news.