Agent Skills
› xuiltul/animaworks
› transcribe-tool
transcribe-tool
GitHub基于Whisper的音频转录工具，支持将音频文件转换为文本。适用于会议录音、播客等场景，具备多语言自动检测及可选LLM后处理能力。
Trigger Scenarios
需要将音频文件转换为文本
处理会议录音或播客的转录需求
Install
npx skills add xuiltul/animaworks --skill transcribe-tool -g -y
SKILL.md
Frontmatter
{
"name": "transcribe-tool",
"tags": [
"audio",
"transcription",
"whisper",
"external"
],
"description": "Audio transcription tool. Converts audio files to text using Whisper, with optional LLM post-processing as needed. Use when: Use when: transcribing meeting recordings, podcast transcripts, or when text extraction from audio files is required."
}
Transcribe Tool
Audio transcription tool using Whisper (faster-whisper).
Invocation
Bash: Run with animaworks-tool transcribe transcribe <音声ファイル> [オプション]
audio — Audio Transcription
animaworks-tool transcribe transcribe audio_file.wav [-l ja] [-m large-v3-turbo]
Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
| audio_path | string | (required) | Path to the audio file |
| language | string | null | Language code (e.g., ja, en). Auto-detected if null |
| model | string | "large-v3-turbo" | Whisper model name |
| raw | boolean | false | If true, skips LLM post-processing |
CLI Usage
animaworks-tool transcribe transcribe audio_file.wav [-l ja] [-m large-v3-turbo]
Notes
- faster-whisper must be installed
- When using GPU, CUDA-compatible ctranslate2 is required
- The model is automatically downloaded on first run
Version History
-
0f2a80e
Current 2026-10-19 11:11
文档国际化更新：生成英文和韩文文档、模板及README，并修复了部分翻译验证问题。
- 140e43b 2026-07-25 09:12
