v0.3.0-beta.1

What's Changed

This release optimizes performance by moving high-frequency data (FFT and call duration) off the main useVoice() context and into dedicated hooks, eliminating cascading re-renders in components that don't need that data.

• Optimize @humeai/voice-react performance: eliminate cascading re-renders by @zachkrall in #417

⚠️ Breaking changes (0.2.x → 0.3.0)

• fft, micFft, and callDurationTimestamp removed from useVoice()
  Use the new hooks instead:

  • usePlayerFft() — assistant audio FFT (replaces useVoice().fft)
  • useMicFft() — microphone input FFT (replaces useVoice().micFft)
  • useCallDurationTimestamp() — formatted call duration (replaces useVoice().callDurationTimestamp)

• FFT types
  usePlayerFft() and useMicFft() return readonly number[]. Update component props from number[] to readonly number[] if you pass these values down.

See the migration guide for before/after examples and full details.

Full Changelog: v0.2.14...v0.3.0-beta.1