Different languages
Each side reads the other in their own language, as it is spoken.
Live translation
You can live on the same street as somebody and share no language either of you is comfortable in. That is true in Tel Aviv, in Berlin, in Toronto and in Dubai, and it is still enough to stop a date from ever happening.
Translating written messages exists everywhere and is not enough. A translated text conversation is slow and distant, and you can hear nothing about the person from it.
What is missing is a conversation. On TOKME you can turn on live captions inside a video or voice call. They are written as you speak, and if the two of you use different languages, each reads the other in their own.
It is not a perfect conversation, and nobody claims it is. But it is a conversation, and that is the difference between meeting somebody and reading a translation of them.
Captions are a feature you switch on. While they are running, short audio segments from the call are sent to our server and passed to a transcription provider to be turned into text.
Those segments are used for transcription only, are not kept by us after processing, and are not used to train models. Switching captions off stops the audio being sent immediately. The call itself, captions or not, runs directly between the devices and is never recorded.
This site is published in English, Hebrew, Arabic, Russian, Spanish, French, Portuguese, German, Italian and Turkish, so understanding what you are downloading does not itself require a second language.
The app interface is currently Hebrew and English, with further languages planned.
Each side reads the other in their own language, as it is spoken.
Captions help just as much when both sides share a language.
Read what you cannot hear, without asking somebody to repeat it again.
Transcription and translation work across a wide range of languages, including English, Hebrew, Arabic, Russian, Spanish and French. Quality varies by language and is affected by background noise and how fast people speak.
No. Audio segments are used for immediate transcription, are not kept after processing, and are not used to train models. Turning captions off stops the sending immediately.
It is good enough for a conversation and not good enough for a contract. The aim is to make possible a conversation that otherwise would not happen, not to replace a human translator.
The app interface is currently Hebrew and English. This site is available in ten languages, and more app languages are planned.
TOKME is in the final stages of store review. Leave an email address and you will be among the first to know on launch day, with a free credit pack for early sign-ups.
The address is used for the launch notice and nothing else. We will not pass it on, and you can leave the list at any time.