openai-whisper-cli

name	openai-whisper-cli
description	Command-line usage for OpenAI Whisper, focusing on audio transcription, translation, and specifying models or languages. You MUST load this skill when interacting with the whisper command.
license	MIT

openai-whisper-cli

WHEN TO USE

• Transcribing audio files into English or other languages using whisper.
• Translating non-English speech into English using whisper --task translate.
• Selecting specific Whisper models (e.g., turbo, medium, large) for speed vs accuracy tradeoffs.
• Specifying the source language for transcription using --language.

WHEN NOT TO USE

• Using OpenAI Whisper from within Python scripts (this skill focuses purely on CLI usage).
• API integrations with OpenAI's hosted Whisper endpoints (this is for local execution).

Core Process

1. Model Selection: Determine the right model. The default is turbo, which is fast and good for English transcription but not for translation. For translating non-English to English, use medium or large.
2. Language Specification: If the audio is not in English, it is often helpful to specify the language with --language.
3. Task Definition: Use --task translate if you need the output in English from a non-English audio file.
4. : Run the command with the appropriate flags and input audio file.