Cantonese (and multi-language) Speech-to-Text for Voice Mode — currently English-only

Problem: Voice Mode's speech-to-text only recognizes English. When I speak Cantonese (or Mandarin) into the mic, it transcribes phonetically-sounding English words instead of Chinese characters — so Voice Mode is unusable for me.

Note: this is the INPUT side (speech recognition), not the TTS output side. The chat model itself understands Cantonese fine; only the mic transcription fails.

Other assistants (ChatGPT, Gemini) already recognize my Cantonese correctly, so this is a clear gap.

Requested:

1. Add Cantonese (and other major languages) to Voice Mode speech-to-text.

2. Ideally auto-detect the spoken language, or let the user pick the input language in Voice Mode settings.

Use case: A user who wants a hands-free conversation in Cantonese without typing or copy-pasting from another app. Right now speaking Cantonese produces garbage, so the only workaround is to type — which defeats the purpose of Voice Mode.

Thank you.