st speak

st-speak — Render a story into spoken audio (MP3)

Renders a story as an MP3 audio file using text-to-speech. Requires the TTS extras to be installed.

Run after: st-prep · Run before: st-post

Examples

st-speak subject.json               # render story 1 → subject.mp3
st-speak -s 3 subject.json          # render story 3
st-speak --source fact subject.json # read the fact-check report aloud
st-speak --voice en_US-ryan-high subject.json  # override voice for this render

Options: -s story --source {text,fact} --voice model -v -q

Related: st-voice · st-post · TTS Audio

---

For developers

Requires pip install 'cross-ai[tts]'. Uses mmd_voice.py which connects to a local Piper TTS server (TTS_HOST/TTS_PORT in .env). Exits cleanly with an error message if TTS dependencies are missing.