메뉴
Fast local speech-to-text on Apple Silicon using MLX Whisper. 10x faster than OpenAI Whisper.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Analyze and process Excel/CSV files using pandas and openpyxl. Supports data summary, filtering, pivot tables, and chart generation.
Playwright browser automation — navigate, read, and interact with web pages using text snapshots and ref-based targeting. Supports keyboard, dialogs, file upload, JS evaluation, console/network debugging, and PDF export. Login state persists across sessions. Use when user wants to open a URL, fill a web form, scrape page content, or operate any website that requires clicking/typing.
Local web search (Tavily/Exa, requires API Key). For quick searches. If no Key configured or deep research needed, use cloud_agent instead.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
Manage Apple Notes via the `memo` CLI on macOS (create, view, edit, delete, search, move, and export notes). Use when a user asks Moltbot to add a note, list notes, search notes, or manage note folders.
Search and manage Apple Photos library on macOS via osascript.
| name | mlx-whisper |
| description | Fast local speech-to-text on Apple Silicon using MLX Whisper. 10x faster than OpenAI Whisper. |
| metadata | {"xiaodazi":{"dependency_level":"external","os":["darwin"],"backend_type":"local","user_facing":true}} |
| capabilities | ["stt","speech_recognition"] |
利用 Apple Silicon 的 MLX 框架本地转录语音，速度是 OpenAI Whisper 的 10 倍。完全离线，隐私安全。
pip install mlx-whisper
需要 Apple Silicon Mac（M1/M2/M3/M4）。
import mlx_whisper
result = mlx_whisper.transcribe(
"audio.mp3",
path_or_hf_repo="mlx-community/whisper-large-v3-turbo",
)
print(result["text"])
result = mlx_whisper.transcribe(
"audio.mp3",
path_or_hf_repo="mlx-community/whisper-large-v3-turbo",
word_timestamps=True,
)
for segment in result["segments"]:
print(f"[{segment['start']:.1f}s - {segment['end']:.1f}s] {segment['text']}")
| 模型 | 大小 | 速度 | 准确度 |
|---|---|---|---|
whisper-tiny | 39M | 最快 | 一般 |
whisper-base | 74M | 快 | 较好 |
whisper-small | 244M | 中 | 好 |
whisper-large-v3-turbo | 809M | 较慢 | 最佳 |
默认使用 large-v3-turbo，短音频（<1分钟）可用 small 加速。
result = mlx_whisper.transcribe("audio.mp3", language="zh")