Openai Whisper Guide

OpenAI Whisper Guide

[中文文档](README_CN.md)

A guide for automatic speech recognition using the optimal Whisper implementation for your hardware. Works with Claude Code and Opencode.

Features

• Environment Detection: Automatically detects your OS, architecture, and GPU to recommend the best Whisper implementation
• Multi-Implementation Support: Covers mlx-whisper, faster-whisper, whisper.cpp, and openai-whisper
• Installation Guides: Platform-specific installation instructions
• Usage Examples: Common transcription patterns and output formats

Quick Start

This skill is designed for use with [Claude Code](https://claude.ai/claude-code) or [Opencode](https://github.com/opencode-ai/opencode). Simply tell Claude:

"Transcribe this audio file"

"Convert video.mp4 to SRT subtitles"

"What's the best Whisper for my Mac?"

"Install whisper and transcribe podcast.mp3"

Or invoke directly with `/openai-whisper-guide`.

# Detect your environment
python3 scripts/detect_env.py

Decision Matrix

Environment	Recommended	Alternative
macOS Apple Silicon	mlx-whisper	whisper.cpp
macOS Intel	whisper.cpp	faster-whisper
Linux + NVIDIA 8GB+	faster-whisper	insanely-fast-whisper
Linux CPU	whisper.cpp	openai-whisper
Windows + CUDA	faster-whisper	whisper.cpp
Windows CPU	whisper.cpp	openai-whisper

Model Recommendation

Use `turbo` model for most tasks—it offers near large-v3 accuracy at 8x speed.

File Structure

openai-whisper-guide/
├── SKILL.md                    # Main documentation
├── scripts/
│   └── detect_env.py          # Environment detection
└── .claude/
    └── settings.local.json    # Bash permissions

References

• OpenAI Whisper
• whisper.cpp
• faster-whisper
• [mlx-whisper](https://githu