Speech to Text
Hey @wang_jianwu, most speech-to-text APIs need you to stick to one language for each transcription session. Speech Lab adds a mixed-language mode (we call this CodeSwitch) that allows you to transcribe audio with a speaker switching between English and Mandarin within the same session. That is, the engine can transcribe casual Singlish-Manadarin speech.
We're also happy to chat if you need a domain-specific language model (that recognises domain-specific terms), or if you want to keep your data off the cloud by deploying locally.
Is there a particular use-case you're looking to support?
I just contribute to the docs for this project :), so I can answer _some_ technical questions but probably not more specific questions.
Can you drop an email at email@example.com? That's a direct line to the product team, who can probably handle your query better.
PS: I've also redacted your phone number from your post (privacy issues) -- if you'd like us to contact you on your phone, please feel free to send it via email to the team. Sorry for the inconvenience!