openai-whisper

Name: openai-whisper
Rating: 78
Author: kody-w

在本地使用 OpenAI 的 Whisper 模型转录音频文件。

SKILL.md

--- frontmatter

name: openai-whisper
description: Transcribe audio files locally using OpenAI's Whisper model.
metadata: {"openclaw":{"emoji":"🎤","requires":{"bins":["whisper"]},"install":[{"id":"pip","kind":"pip","module":"openai-whisper","bins":["whisper"],"label":"Install whisper (pip)"}]}}

OpenAI Whisper (Local)

Transcribe audio files using the local Whisper model.

Basic Transcription

bash

whisper audio.mp3 --model base --output_format txt

With Language Detection

bash

whisper audio.mp3 --model medium --task transcribe --output_format json

Translation to English

bash

whisper foreign-audio.mp3 --model medium --task translate

Models

Model	Size	Quality
tiny	39M	Basic
base	74M	Good
small	244M	Better
medium	769M	Great
large	1550M	Best