[codex] Add capture-only realtime and stream playback

[rgbkrk]

Summary

Follow-up to the merged realtime voice loop PR. This adds the two runtime pieces Vodex needs for the current local voice adapter:

• allow voice realtime --mic --json to run transcript-only without --speaker
• add voice stream --play so daemon-streamed TTS frames can play through the foreground CLI as they arrive

Why

Vodex now uses voice realtime --mic --json as its live STT capture surface and voice stream --play as its warm daemon-backed TTS playback surface. This keeps the Python bridge out of the audio path while still letting the Codex-facing prototype use streaming mic capture and streaming TTS.

Validation

• cargo check -p voice
• cargo test -p voice realtime_live -- --nocapture
• cargo test -p voice stream -- --nocapture
• cargo fmt --check -p voice

Notes

This PR is intentionally scoped to the CLI/runtime surfaces required by Vodex. It does not add the Vodex prototype itself to this repository.

[pullfrog]

This run was cancelled 🛑

The workflow was cancelled before completion. Please check the link below for details.

  ^{｜ View workflow run ｜ via Pullfrog ｜ 𝕏}