MJ Windows Faster Whisper

该 Skill 使用 faster-whisper（CTranslate2）在本地把音频转成纯文本，包含安装 Python 依赖、准备 ffmpeg、下载并复用模型缓存的步骤；默认 CPU int8，不适用于 whisper.cpp 的 ggml 模型。

热度 41.90

最近更新 2026-05-11 17:45:33

分类：效率依赖：代码执行 / 读写文件 / 第三方 API应用场景：文件转换 / 数据处理 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add mj-windows-faster-whisper

Metadata

name: faster-whisper
description: Local speech-to-text with the faster-whisper backend (CTranslate2). Use when transcribing audio locally, setting up the faster-whisper model cache, or replacing a whisper-cli workflow with a faster local engine.

Use faster-whisper for local transcription with low latency and a reusable model cache.

Do not assume ggml models work here; faster-whisper uses CTranslate2 model folders.
Prefer CPU device='cpu' and compute_type='int8' unless the machine is explicitly configured for GPU.
Keep output plain text unless the user asks for timestamps or captions.

Confirm python and ffmpeg are available.
Install the Python packages needed for local inference:
- faster-whisper
- ctranslate2
- huggingface_hub
Use the project repo https://github.com/SYSTRAN/faster-whisper for install/setup guidance.
Download Systran/faster-whisper-small from https://huggingface.co/Systran/faster-whisper-small into a stable local folder such as:
- C:\Users\joshu\.openclaw\tools\faster-whisper\models\Systran-faster-whisper-small
Reuse that folder for repeat runs.
If the user only has a ggml-*.bin file, explain that it belongs to whisper.cpp and is not usable here.