WhisperKit Voice

Voice for your agent, no API required.

Features: Text-to-speech with 9 built-in voices, speech-to-text with multiple model sizes, speaker diarization to identify who spoke when, style instructions for voice personality, and OpenAI-compatible local server.

Why WhisperKit

vs ElevenLabs: Free vs $99-299/month, ~150ms vs ~300ms latency, on-device vs cloud vs Chatterbox: No auth vs HuggingFace gated, native Apple vs Python, STT+TTS vs TTS only

Voices

aiden - Clear, professional ryan - Warm, natural sohee - Gentle, calm eric - Deep, authoritative

Setup

brew install whisperkit-cli

Attribution

Uses Argmax WhisperKit (MIT)