What maintained, openly licensed Flutter Android example would you recommend for an opt-in foreground voice assistant with no wake word, conversational speech, and reliable interruption while narration is playing?
The target interaction is: start a voice session, ask a question normally, hear the answer, say Stop or Pause during speech, then ask another question or resume narration. Recognition should tolerate hesitations and short commands. The microphone must remain usable during output.
I am looking for free community advice about:
- A minimal full-duplex Flutter/WebRTC sample with echo cancellation, Android audio-focus handling and clean session teardown.
- A diagnostic checklist separating microphone capture, turn detection, asynchronous transcription ordering, command dispatch and audible playback-stop latency.
- A public example of word-by-word read-along tied to actual playback: only the current word becomes highlighted/bold; the preceding word returns to its original style without shrinking the text.
- A sensible test matrix for headset versus speaker output, including a user deliberately repeating words spoken by the assistant.
Links to specific example files or documentation are especially helpful. This is a request for public reference code and methods, not a claim of a Flutter SDK defect or a request for paid development. Thank you.