feat(voice): add OpenAI text-to-speech

[johnmatthewtennant]

Summary

Adds OpenAI text-to-speech as a Voice Conversation output backend.

• Adds an OpenAI API-key field in Voice settings.
• Stores the key as Berd's shared OpenAI voice credential and never reads Goose provider credentials.
• Streams synthesized speech through the selected output device with interruption and delivery tracking.
• Stops active playback when the credential changes or is removed.
• Localizes setup guidance and distinguishes missing TTS prerequisites.

Configuration

Enter an OpenAI key in Settings → Voice → Speech output after selecting OpenAI. Berd stores it in the system credential manager under the shared OpenAI voice entry. Goose provider credentials remain separate.

[morgmart]

🤖 Automated code review

APPROVE. Fresh static review completed for the exact three-dot comparison 771776f...382df01. The code-review and wes-review passes covered the OpenAI TTS credential setup, readiness and settings UI, backend selection, streaming synthesis/playback, interruption and delivery accounting, window teardown, persistence, platform behavior, localization, and tests. Final self-check covered every changed user flow; accessibility; English and Spanish localization completeness; navigation and consent guards; async failure, never-completes, lifecycle, and race behavior; test honesty; design-system and project rules; duplicate overlap; and the evidence/user-effect bar for blocking severity. No publishable findings survived. Supplied GitHub evidence was inspected: all ten reported check runs completed successfully; the separate combined commit-status state is pending and required checks still govern merge readiness.

Deterministic publication result: 0 blocking and 0 non-blocking finding(s) publishable; 0 duplicate(s) suppressed.

Pending checks: 1 check(s) are not complete.

This approval reflects the completed code review only; merge readiness remains governed by the repository's required checks.

[morgmart]

🤖 Automated code review

REQUEST_CHANGES. Fresh static review completed for the exact three-dot comparison 771776f...b1268c3. The code-review and wes-review passes covered credential setup and migration, readiness and settings UI, backend selection, streaming synthesis and playback, interruption and delivery accounting, window teardown, persistence, platform behavior, localization, and changed tests. Final self-check covered every changed user flow; accessibility; English and Spanish localization completeness; navigation and consent guards; async failure, never-completes, lifecycle, and race behavior; test honesty; design-system and project rules; duplicate and overlapping findings; and concrete evidence plus user effect for blocking severity. One blocking credential-state consistency issue remains. Supplied GitHub evidence was inspected: all ten reported check runs completed successfully; the separate combined commit-status state is pending, and required checks still govern merge readiness.

Deterministic publication result: 1 blocking and 0 non-blocking finding(s) publishable; 0 duplicate(s) suppressed.

[morgmart]

🤖 Automated code review

APPROVE. Fresh static review completed for the exact three-dot comparison 96a91d9...02b5005. The code-review and wes-review passes covered credential setup and migration, readiness and settings UI, backend selection, streaming synthesis and playback, interruption and delivery accounting, window teardown, persistence, platform behavior, localization, and changed tests. The previously reported credential-state inconsistency is fixed: canonical mutation now determines success while legacy cleanup is best-effort, with regression coverage. Final self-check covered every changed user flow; accessibility; English and Spanish localization completeness; navigation and consent guards; async failure, never-completes, lifecycle, and race behavior; test honesty; design-system and project rules; duplicate and overlapping findings; and concrete evidence plus user effect for blocking severity. No publishable findings survived. Supplied GitHub evidence was inspected: all ten reported check runs completed successfully; required checks still govern merge readiness.

Deterministic publication result: 0 blocking and 0 non-blocking finding(s) publishable; 0 duplicate(s) suppressed.

Pending checks: 1 check(s) are not complete.

This approval reflects the completed code review only; merge readiness remains governed by the repository's required checks.