Media TTS AI

Context: $ARGUMENTS

Quick start

Default TTS (no cloning), best quality-for-size: scripts/tts.py speak --model kokoro --text "Hello" --voice af_bella --out out.wav → Step 3
Voice clone from a 5-10s sample: scripts/tts.py clone --model openvoice --reference ref.wav --text "..." --out out.wav → Step 4
Audiobook from a text/markdown file: scripts/tts.py audiobook --text-file book.md --model kokoro --voice af_bella --chapter-split --out-dir chapters/ → Step 5
Realtime embedded (lowest latency, local): scripts/tts.py speak --model piper --text "Hi" --voice en_US-amy-medium --out out.wav
Expressive + non-verbal ([laughs], [sighs]): scripts/tts.py speak --model bark --text "Nice. [laughs]" --out out.wav