Real-Time Voice Translation
A real-time voice translation app between user and human-interlocutor where user selects user's language and then user click button to start the conversation.
Core trick
Streams interlocutor audio through LiveKit into Google real-time speech translation, returning translated audio and text with three suggested replies.
Tags
Description
A real-time voice translation app between user and human-interlocutor where user selects user's language and then user click button to start the conversation.
When user gets audio input from interlocutor on interlocutor's language then app should translate this audio input to the user's language and output to the user as audio and text.
App prepares 3 different options on how user can reply on the interlocutor's input and then app outputs result as text using source language to the user.
Also when app gets text or audio input from the user then app translates this user's input to the interlocutor's language, outputs as text to the user in interlocutor's language and if needed tries to answer in text on this question using user language.
Based on LiveKit and Google Real-time speech sdk https://docs.livekit.io/agents/integrations/google
The linked repo returns 404, so it was deleted or made private.