AgentSkillsCN

audiobook

通过 Google Gemini TTS 与 Vertex AI 将 Calibre 图书馆中的书籍章节转换为有声读物（MP3 格式）。以专业级语音提示呈现高品质的有声朗读内容。

SKILL.md
--- frontmatter
name: audiobook
description: Convert book chapters from Calibre library to audiobook (MP3) using Google Gemini TTS via Vertex AI. High-quality audiobook narration with professional voice prompts.
allowed-tools: Bash(calibredb:*) Bash(ebook-convert:*) Bash(uv:*) Bash(rm -rf /var/folders:*) Bash(rm -rf /tmp:*)
argument-hint: <chapters> книги <book title>
metadata:
  author: aleksandrbobrov
  version: "3.0"
compatibility: Requires Calibre, ffmpeg, uv, Google Cloud auth (gcloud auth application-default login)

Audiobook Creator Skill

Converts specific chapters from books in your Calibre library to MP3 audiobooks using Google Gemini TTS.

Example Commands

  • •/audiobook 7 и 8 главу книги Убийства и кексики
  • •/audiobook chapters 1-3 of "The Great Gatsby"
  • •/audiobook главы 5, 6, 7 книги "Война и мир"

Configuration

Calibre Library Path: /Users/aleksandrbobrov/Library/Mobile Documents/com~apple~CloudDocs/CalibreLib

Scripts Directory: scripts/ (relative to this SKILL.md)

Workflow

Step 1: Parse User Request

Extract:

  • •Book title: e.g., "Убийства и кексики"
  • •Chapters: e.g., "7 и 8" → chapters 7, 8

Chapter formats:

  • •главу 7 / chapter 7 → single chapter
  • •главы 7 и 8 / chapters 7 and 8 → multiple
  • •главы 1-5 / chapters 1-5 → range
  • •главы 1, 3, 5 → specific list

Step 2: Find Book in Calibre

bash
CALIBRE_LIB="/Users/aleksandrbobrov/Library/Mobile Documents/com~apple~CloudDocs/CalibreLib"
calibredb list --library-path "$CALIBRE_LIB" --search "title:~<book_title>" -f title,authors,formats --for-machine

Get book ID from results.

Step 3: Export and Convert

bash
WORK_DIR=$(mktemp -d)
calibredb export <BOOK_ID> --library-path "$CALIBRE_LIB" --to-dir="$WORK_DIR" --single-dir --dont-write-opf --dont-save-cover --formats=epub,fb2

# Convert to text
BOOK_FILE=$(ls "$WORK_DIR"/*.{epub,fb2,mobi} 2>/dev/null | head -1)
ebook-convert "$BOOK_FILE" "$WORK_DIR/book.txt"

Step 4: Extract Chapters

bash
SCRIPTS_DIR="<path-to-this-skill>/scripts"
uv run --project "$SCRIPTS_DIR" python "$SCRIPTS_DIR/extract_chapters.py" \
  "$WORK_DIR/book.txt" \
  --chapters 7,8 \
  --output "$WORK_DIR/chapters.md"

To list chapters first:

bash
uv run --project "$SCRIPTS_DIR" python "$SCRIPTS_DIR/extract_chapters.py" "$WORK_DIR/book.txt" --list

Step 5: Convert to Audiobook

bash
uv run --project "$SCRIPTS_DIR" python "$SCRIPTS_DIR/md_to_audiobook.py" \
  "$WORK_DIR/chapters.md" \
  --output ~/Downloads/"<BookTitle>_chapters_<X-Y>.mp3"

Step 6: Cleanup

bash
rm -rf "$WORK_DIR"

Voice

Uses Despina voice - warm, intimate narration style optimized for audiobooks.

Response Format

code
Found book: "<Title>" by <Author>
Book ID: <ID>

Extracting chapters <X-Y>...
Content: ~<N> characters

Converting to audiobook...
Voice: <voice>

Audiobook saved to: ~/Downloads/<filename>.mp3
Size: <X.XX> MB

Error Handling

  • •Book not found: Show available books with similar titles
  • •Chapter not found: List available chapters with --list
  • •Auth error: Run gcloud auth application-default login
  • •ffmpeg missing: "Install with: brew install ffmpeg"