Microphone Function Cannot Process Message

Bug: Speaking practice feature fails with ASR_ERROR

Hi, I’ve been trying to use the speaking/pronunciation practice feature for Chinese (zh-CN) and it consistently fails. I’ve done a lot of debugging and wanted to share what I found.

Environment

  • Browser: Chrome 148.0.7778.215 (Official Build) (64-bit)
  • OS: Linux Mint
  • Extension: Language Reactor only (tested in a brand new Chrome profile with no other extensions installed)
  • Target language: Chinese (zh-CN)

What happens
I start hitting record, I record my message, I release the mic button. I see the audio in the chat line. I hit send, and I get an error: “We couldn’t process your message. Please try again.”
It’s important to note that when I record my voice and stop, I get the following error in the Chrome DevTools console before anything is even sent to the server (before I hit the send button):

endRecordingAudio is called when isBusy is false, nothing to do.

When the request does reach the server, the API call to https://api-cdn-plus.dioco.io/base_chat_getUserPrompts returns:

{ "status": "failure", "error": "ASR_ERROR" }

What I’ve already tried/ruled out

  • Microphone permissions: confirmed allowed for languagereactor.com in Chrome settings
  • Microphone hardware: works correctly on all other sites
  • Web Speech API: confirmed working for both English and Chinese (zh-CN) on Google’s own demo at Chrome Browser
  • Other extensions: reproduced in a brand new Chrome profile with Language Reactor as the only installed extension
  • Chrome language settings: Chinese languages added under chrome://settings/languages
  • Chrome flags: reviewed audio-related flags, nothing obviously misconfigured

I’ve run out of things to try on my end. Happy to provide any additional logs or run any tests if it helps. Thanks.