@cloudflare/voice-deepgram@0.1.0 @cloudflare/voice-elevenlabs@0.1.0

low risk

microsoft/VibeVoiceAug 27, 2026 → Aug 27, 20260 days between releases

What's New

Switches provider from Deepgram to ElevenLabs with identical lifecycle improvements.

New Features

  • Clear stale interim transcripts during startup
  • Add structured, content-free browser diagnostics
  • Report transcriber startup and runtime failures
  • Distinguish no-output, output-limit, and model-error completions
  • Add stable typed per-turn timing summaries

Improvements

  • Enhance voice lifecycle accuracy and diagnostics
  • Improve error reporting and call cleanup
  • Add comprehensive timing summaries for streaming and reasoning
  • Keep existing four-field metrics wire shape compatible
  • Update bundled voice providers to propagate lifecycle failures

Upgrade Notes

The package name changed from `voice-deepgram` to `voice-elevenlabs`. No API changes are required as the implementation and changelog are identical. Ensure you update your import paths.

The Two Releases

@cloudflare/voice-deepgram@0.1.0

@cloudflare/voice-deepgram@0.1.0

Aug 27, 2026· 2 months ago

A minor update focusing on lifecycle accuracy, diagnostics, and per-turn timing visibility for Deepgram integration.

Full release notes
@cloudflare/voice-elevenlabs@0.1.0

@cloudflare/voice-elevenlabs@0.1.0

Aug 27, 2026· 2 months ago

Introduces improved lifecycle accuracy, diagnostics, and per-turn timing visibility for ElevenLabs voice generation.

Full release notes