How it works
From your voice to theirs — and back.
Kanutus sits between your microphone and your call app. You talk normally; the other side hears you in their language, in your voice. Here is the whole path.
Tom (you)EN
AnaPT
MarcusEN
LenaFRCreate your voice
Open the web app and read a short text for 30 to 60 seconds. Kanutus builds a voice that keeps your timbre and rhythm. Prefer not to? Pick one of the ready-made voices instead. Voices can be re-recorded or deleted at any time.
Choose the languages
Tell Kanutus what you speak and what the room speaks. Translation runs in both directions: what you say goes out translated; what they say comes back to your headphones in your language.
Join the call
Start Kanutus from the tray and select the Kanutus microphone in your call app. Hold a key to push-to-talk, or leave it hands-free. A small floating island shows what you said and what went out, so you always know what the room heard.
Share the room link
Anyone without Kanutus can open your room link in a browser — desktop or phone — pick a language and hear the call in it. No account, nothing to install.
Under the hood
Three stages, about a second apart.
Each sentence you finish goes through the pipeline on its own, so translation keeps pace with the conversation instead of waiting for you to stop talking.
Listen
Speech recognition
Your microphone audio is transcribed as you speak, sentence by sentence.
Translate
Meaning, not words
A translation model built for conversation keeps idioms, numbers and tone intact — "beat a dead horse" becomes the expression a native would use.
Speak
Your voice
The translated sentence is spoken in your cloned voice and sent into the call as your microphone.
Both directions
You hear them in your language too.
Kanutus listens to the call audio and plays the translation in your headphones. Speaking and listening are tracked separately, and listening counts only partially towards your minutes.
You → the room
Your sentence, translated, spoken in your voice through the virtual microphone.
The room → you
What they say, translated, in your headphones a moment later.
Room link
One link for everyone else.
A browser page on kanutus.com where each listener picks their own language. Works on phones. Great for a prospect, a candidate or a patient who will never install anything.
What you need
A computer and a headset.
- Windows 10 or 11, macOS 12 or later, or Linux (Ubuntu 22.04+, Fedora, Debian)
- A headset or earphones, so the translation you hear doesn't feed back into the call
- A stable internet connection — processing runs on our servers
- Any call app that lets you choose a microphone
Audio is processed, not kept.
Call audio is used only to transcribe, translate and speak during the session. It is not stored afterwards and is never used to train models. Your voice sample stays yours and can be deleted at any time.
More questions
01Can I use it without cloning my voice?
Yes. Pick a ready-made voice in the web app. You can switch to your own voice later.
02What do the other people on the call see?
Nothing new. Kanutus doesn't join the meeting as a participant or bot. Your video and your name stay as they are; only the audio of your microphone is translated.
03Does it work on phone calls?
On your computer, any app with a microphone setting works — WhatsApp Desktop, Skype, Discord. Calls placed from a phone handset are not supported yet.
04Can several people on my side use it?
Each speaker needs their own Kanutus account and voice. Teams of five or more usually fit the Business plan; larger ones talk to us.
Ready to try it on a real call?
Download Kanutus, create your voice, pick your languages and join your next meeting.