Voice dictation: support multiple simultaneous languages

Feature Request

Description:
Support configuring multiple languages for voice dictation simultaneously, rather than only one at a time.

Use case:
Bilingual users (e.g., Portuguese + English) frequently switch between languages mid-conversation. Currently, changing the dictation language requires manually going to /config each time, which breaks the flow.

Proposed solution:
Allow the language setting to accept multiple language codes (e.g., ["pt", "en"]), enabling the speech recognition engine to auto-detect which language is being spoken.

Current behavior:
Only one language can be set at a time via /config.

Expected behavior:
Voice dictation recognizes and transcribes speech in any of the configured languages without manual switching.