Summary
When using voice dictation (Windows Speech Recognition or similar) to compose messages in the Amazon Quick desktop app, switching focus away from the Quick window while the assistant is streaming a response causes the response to stop mid-stream. The response does not resume when focus returns — it appears truncated/incomplete.
Steps to Reproduce
- Open the Amazon Quick desktop app (Windows)
- Start a conversation and ask the assistant a question that produces a long streaming response
- While the response is still streaming, activate voice dictation (e.g., Win+H or another speech input method)
- The dictation overlay/window steals focus from the Quick app
- Observe that the response stream stops immediately
Expected Behavior
The assistant’s response should continue streaming to completion regardless of whether the application window has focus. Background streaming should not depend on window focus state.
Actual Behavior
The response stream is interrupted when the window loses focus. The partial response remains visible but is incomplete, and there is no way to resume or regenerate the rest of the response without re-asking the question.
Environment
- App: Amazon Quick Desktop (Windows)
- OS: Windows 11
- Trigger: Voice dictation (Win+H) or any action that moves focus away from the Quick window during response streaming
Additional Notes
This is particularly disruptive for accessibility workflows that rely on voice input, since dictation inherently requires focus changes. Users who depend on voice input cannot reliably use the app while responses are streaming.