A comprehensive local pipeline for turning raw text into multi-character audiobooks using intelligent AI casting and StepFun TTS technology.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install step-audiobook
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install step-audiobook using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Audiobook Skill?
The audiobook skill is a sophisticated production environment for developers looking to build and manage voice libraries while automating the synthesis of complex audio content. As a key part of the Openclaw Skills collection, it provides a 4-stage workflow covering voice synchronization, script structuring, character casting, and segment-based synthesis. This tool is designed to work locally, giving users full control over intermediate artifacts and the ability to refine AI-driven casting decisions before committing to final audio generation.
Built to leverage StepFun specialized audio models, the skill handles the heavy lifting of voice character analysis and long-context script understanding. It is particularly useful for creators who need a repeatable, auditable process for generating high-quality voice content without relying on opaque, one-click solutions. By using Openclaw Skills, you can ensure that every step of the audiobook lifecycle—from character profiling to final MP3/WAV export—is manageable via professional CLI tools.
Audiobook Skill Use Cases
- Automating the conversion of long-form TXT novels into structured, multi-voice dramatized scripts.
- Building and maintaining a local repository of official and cloned voice assets for consistent branding.
- Performing intelligent LLM-driven casting to match fictional characters with specific vocal profiles.
- Regenerating specific audio segments (TTS replay) without needing to re-synthesize the entire book.
- Previewing and reviewing AI-generated casting plans before performing paid voice cloning operations.
How Audiobook Skill Works
- The skill synchronizes the local voice library and analyzes reference audio files using Step-Audio models to create vocal profiles.
- Raw text is parsed into a structured script format, identifying dialogue, narration, and character traits.
- An AI reasoning layer (defaults to Step-3.5) generates a casting plan, matching characters in the script to available voices in the library.
- Replayable TTS requests are generated in JSON format, allowing for granular control over synthesis parameters.
- Individual audio segments are synthesized and then merged using ffmpeg into a final high-fidelity audiobook file.
Audiobook Skill Setup
To get started with this skill from the Openclaw Skills library, ensure you have the necessary dependencies and environment variables configured:
# Install system dependencies
brew install ffmpeg ffprobe
# Configure your environment
export STEP_API_KEY="your_stepfun_api_key"
# Initialize your local voice library
mkdir -p ~/.openclaw/workspace/audiobook-library
cp assets/voice-library.template.yaml ~/.openclaw/workspace/audiobook-library/voice-library.yaml
# Run the complete workflow
python3 scripts/run_audiobook.py --library "$VOICE_LIBRARY" --story-input /path/to/story.txt
Audiobook Skill Data Schema & Taxonomy
The skill organizes data using a structured directory hierarchy and specific YAML/JSON schemas to ensure interoperability:
| File Type | Description |
|---|---|
voice-library.yaml |
The master database of voice assets, profiles, and clone configurations. |
structured-script.yaml |
The intermediate representation of the text, broken down by character and chapter. |
casting-plan.yaml |
A mapping file that assigns specific voices to characters based on AI recommendations. |
artifacts.json |
A manifest file generated for every run to track all intermediate and final outputs. |
tts-requests.json |
A batch list of synthesis requests used for generating audio segments. |
Audiobook Skill Advanced Features
- Multi-stage human-in-the-loop support allowing manual edits to scripts, casting, and voice reviews.
- Secure paid-action hooks requiring explicit confirmation flags before triggering costly voice cloning APIs.
- Flexible LLM reasoning layer that can be redirected to any OpenAI-compatible endpoint for script analysis.
- Automated segment-based synthesis with smart caching to reduce API costs and improve iteration speed.
- Detailed trace and debug logging within a dedicated local trace directory for every story run.
SKILL.md
Loading
Related Openclaw Skills
Skill Vetter
A mandatory security-first vetting protocol that inspects and auto-blocks high-risk AI agent skills before installation.
Skill Router Pro
Skill Router Pro is a dedicated discovery assistant that manages your library of Openclaw Skills by searching local indexes and ClawHub before you build new tools.
China Job Search
A specialized tool for aggregating and filtering job listings from China's top recruitment platforms including BOSS Zhipin, Zhilian, and 51job.
Hono Framework Skill
Hono is a lightweight, blazing-fast web framework built on Web Standards that operates seamlessly across any JavaScript runtime.
TikTok Shop Fulfillment
Expert guidance for TikTok Shop fulfillment, including setup for Fulfilled by TikTok, self-fulfillment, shipping templates, and return policies.
Qiushi Openclaw Skills
A collection of strategic methodology tools for AI agents designed to provide a systematic framework for thinking and logical action.
Featured*
