vercel/ai @ai-sdk/elevenlabs@2.0.70

on GitHub

Patch Changes

• 313e084: fix(elevenlabs): map the generic HTTP speech output format wav to the native wav_44100 format instead of forwarding an unsupported value.
• ceaafcb: Add the withTimestamps speech option to return character alignment and single-voice Eleven v4 voice segments in provider metadata. Preserve Scribe batch word and character timestamp metadata.
• Updated dependencies [28252da]
  • @ai-sdk/provider-utils@4.0.63