| name | epub-to-vibevoice |
| description | Convert EPUB books into clean, narration-friendly plain text for local VibeVoice audiobook generation, with optional chunking for stable long-form TTS. Use when the user has an .epub and wants text cleanup, chunk prep, or a local audiobook workflow for VibeVoice. |
| tools | Bash |
| risk | low |
| source | local |
| date_added | 2026-05-02 |
EPUB -> VibeVoice Text Prep
Prepare EPUB books into pure, narration-friendly text for VibeVoice.
Use this skill when
- You have an
.epub and want local TTS/audiobook generation.
- You need cleaner text (remove markdown/noise/watermarks) before synthesis.
- You want optional chunk files for stable long-form generation.
Commands
1) Convert EPUB to clean text
cd ~/VibeVoice
source venv/bin/activate
python scripts/epub_to_vibevoice_text.py \
--input "/path/to/book.epub" \
--output "/path/to/book.clean.txt" \
--start_regex "^(Chapter|CHAPTER|Prologue|PROLOGUE|To you\.|When Red wins\.)"
2) Convert + chunk for TTS
cd ~/VibeVoice
source venv/bin/activate
python scripts/epub_to_vibevoice_text.py \
--input "/path/to/book.epub" \
--output "/path/to/book.clean.txt" \
--chunks_dir "/path/to/chunks" \
--chunk_chars 1800 \
--start_regex "^(Chapter|CHAPTER|Prologue|PROLOGUE|To you\.|When Red wins\.)"
3) Run VibeVoice audiobook generation on clean text
cd ~/VibeVoice
source venv/bin/activate
python scripts/generate_audiobook_en.py \
--input "/path/to/book.clean.txt" \
--output "/path/to/book.wav" \
--voice en-soother_woman \
--cfg_scale 1.2 \
--device mps
Notes
chunk_chars: 1200–2200 is usually stable.- Keep punctuation and paragraph breaks for better prosody.
- This workflow is optimized for English narration with VibeVoice Realtime.
References
