Settings¶

Settings go in the front matter of script.md. Everything is optional; defaults live in tools/studio/common.py. Changing a voice setting re-synthesises the narration.

---
title: My Video
tts: kokoro
voice: bf_emma*0.6 + bf_isabella*0.4
speed: 0.85
---

Episode¶

Setting	Default	Meaning
`title`	the slug	Episode title
`format`	`landscape`	`landscape` (1920×1080) or `reel` (1080×1920). See Reels
`burn_captions`	`false`	Burn word-highlighted captions into the video
`fps`	`30`	Frame rate
`pronounce`	none	TTS-only respellings, such as `{ Engels: Engles }`

Voice¶

Setting	Default	Meaning
`tts`	`kokoro`	`kokoro`, `chatterbox` or `elevenlabs`
`voice`	`bf_emma0.6 + bf_isabella0.4`	Kokoro voice or blend; ElevenLabs voice ID; Chatterbox clip path or `default`
`speed`	`0.85`	Speaking rate, relative to each voice. ElevenLabs accepts 0.7–1.2; Chatterbox ignores it

Kokoro¶

Setting	Default	Meaning
`g2p`	`misaki`	Phonemizer: `misaki` or `espeak`
`sentence_pause`	`0.38`	Seconds of silence between sentences
`clause_pause`	`0.12`	Seconds of silence where an over-long sentence is split at a comma
`voice_fx`	`true`	Gentle EQ, de-ess, compressor and faint room

Chatterbox¶

Setting	Default	Meaning
`chatterbox_model`	`turbo`	`turbo` (350M) or `standard` (500M)
`exaggeration`	`0.5`	`standard` only. Emotional intensity, 0.25–2
`cfg_weight`	`0.5`	`standard` only. Lower is slower and more deliberate
`temperature`	`0.8`	Sampling temperature
`seed`	`0`	Random seed, so a re-synthesis repeats

sentence_pause, clause_pause and voice_fx apply to Chatterbox as well.

ElevenLabs¶

Setting	Default	Meaning
`tts_model`	`eleven_multilingual_v2`	ElevenLabs model
`voice_settings`	none	`{stability, similarity_boost, style, use_speaker_boost}`

Timing¶

Setting	Default	Meaning
`pad_in`	`0.6`	Silence before a scene's narration starts
`pad_out`	`0.9`	Silence after it ends
`paragraph_gap`	`0.45`	Pause between paragraphs inside a scene

Music and loudness¶

Setting	Default	Meaning
`music`	none	Path to a music bed, relative to the studio root
`music_prompt`	none	Default prompt for `studio music`
`music_db`	`-24`	Level under narration, before ducking
`music_fade_in`	`2`	Seconds. `0` keeps a hit on the bed's first beat
`loudness`	`-14`	Integrated LUFS target
`true_peak`	`-1.5`	dBTP ceiling

Environment¶

Variable	Meaning
`ELEVENLABS_API_KEY`	ElevenLabs key. Also read from `.env` at the studio root (`studio setup` writes it)
`STUDIO_ARCHIVE`	rclone path for `studio archive`. Also read from `.env` (`studio setup` writes it)
`STUDIO_HF_VERSION`	Override the pinned HyperFrames version
`STUDIO_CHATTERBOX_VERSION`	Override the pinned Chatterbox version