[FEATURE] Support multiple input languages in the STT of Agora Extension

[zshbleaker]

Description

In multiple use cases the user may speak multiple languages. Being able to detect the language the user is speaking is essential for the correct transcription.

Currently up to 2 languages is supported in Agora Real-Time Speech to Text. The Agora Extension in TEN should be aligned as well. https://docs.agora.io/en/real-time-stt/get-started/quickstart

Severity

Major

Additional Information

No response