Voice-to-text appears in wrong workspace chat when switching during TTS loading

Support Bug Reports
, ,
1

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

When using voice input in workspace1, during the voice-to-text conversion (TTS read-out loading phase), switching to workspace2 before conversion completes causes the transcribed text to appear in workspace2ā€™s chat input field instead of workspace1.

The issue is that the display location of the converted text is wrong.

Workaround: wait for voice-to-text conversion to finish in workspace1, confirm the text is reflected there, then switch to workspace2.

Steps to Reproduce

  1. Open two workspaces (workspace1 and workspace2) in Cursor
  2. In workspace1, open agent chat and start voice input
  3. Speak and finish recording
  4. While voice-to-text conversion / TTS loading is still running, switch to workspace2
  5. Observe: transcribed text appears in workspace2ā€™s chat input (not workspace1)

Expected Behavior

Voice-to-text conversion started in workspace1 should always display the transcribed text in workspace1ā€™s chat input, even if the user switches to another workspace during the TTS loading phase.

Screenshots / Screen Recordings

Operating System

Windows 10/11

Version Information

OS: Windows 10.0.26200
(Please fill in Cursor IDE version from Menu ā†’ About Cursor ā†’ Copy before posting)

Additional Information

Reproduction video will be attached before posting.

Workaround: wait for voice-to-text conversion to complete in workspace1, confirm text is reflected, then switch to workspace2.

Enter is not pressed during the bug, so the agent is not affected ā€” only the chat input display location is wrong.

Does this stop you from using Cursor

No - Cursor works, but with this issue

5

Hey @ll1!

Thanks for reporting this. We had the exact same report internally, and Iā€™m linking your report to that existing bug report.