[waleedlatif1]

Summary

• Make the microphone control stop voice input while its token, permission prompt, or provider connection is still pending.
• Release the microphone immediately when startup is cancelled and ignore late results, while preserving final transcripts after an established session stops.
• Extend packaged browser coverage with a delayed provider connection and token response, observing real microphone tracks through cancellation and restart.

Type of Change

• Bug fix

Testing

Scoped Biome and git diff --check passed. Pre-fix CI fails as intended: after cancellation, the audio track remains live and the connection stays pending. Fixed-head CI passed: full lint/type/audit, unit/integration, production builds, desktop/mobile, and hosted HTTP/browser acceptance. Packaged transport:32 passed; extension navigation:10; composer:16 plus1 optional live-provider skip; access:16. Browser errors:0. Both reviewers:5/5, zero open threads. Provider transport is simulated for deterministic cancellation coverage; this does not claim live-provider acceptance.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Nov 5, 2026 2:04am UTC

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable defects were found.

Summary

Makes voice startup cancellable while the token request, microphone permission, or provider connection is pending.

• Voice input can now be stopped while startup is still waiting.

No actionable issues were found. Review was based on source inspection; browser checks were not run.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Click microphone] --> B[Show stop control]
  B --> C[Request token]
  C --> D[Request microphone]
  D --> E[Start capture timer]
  E --> F[Connect to provider]
  F --> G[Stream audio]
  B --> H[Cancel startup]
  C --> H
  D --> H
  F --> H
  H --> I[Invalidate attempt and release resources]
  G --> J[Stop established session]
  J --> K[Release microphone and request final text]
  K --> L[Close connection after grace period]

Loading

_{Reviews (1) · Last reviewed commit: "fix(voice): stop microphone capture duri..." · Reviewed by Greptile}

[cubic-dev-ai]

No issues found across 2 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Turn on auto-fix | Re-trigger cubic}