# Audio Transcribe

- Name: Audio Transcribe
- Collection: Simon personal skills
- Origin: Simon workflow and wrapper; upstream tools and models are separate
- Author: Simon Lin / Simon 林
- Copyright: PhoenixTree Holdings Limited
- Distribution version: 1.1.0
- Original package snapshot: 2026-08-18
- Free distribution published: 2026-10-22 (Asia/Singapore)
- Documentation and distribution checked: 2026-10-22
- License for original contents: MIT
- Verification: documentation_and_distribution_checks_only; full model or browser workflows not rerun
- Download: https://agentskill.nz/downloads/audio-transcribe-v1.1.0.zip
- Bytes: 6968
- SHA-256: 177e7bd53d082fb188481ea2c867bd1f35adf273d1129bc939ceebd9c1e88e92
- Chinese page: https://agentskill.nz/personal/audio-transcribe/
- English page: https://agentskill.nz/en/personal/audio-transcribe/

## Upstream components

- OpenAI Whisper: https://github.com/openai/whisper (MIT); not bundled

## 中文

本地音频转文字

整理 OpenAI Whisper 的本地音频转录、批处理与字幕工作流。

前提：Python、ffmpeg、openai-whisper；首次使用需单独下载模型。

- 单文件与字幕说明
- 批处理和失败记录方法
- 音质与人工校对边界

此包为工作流文档；按 SKILL.md 准备环境并核对 whisper --help 后处理自己的媒体。

## English

Local audio transcription

A local audio transcription, batch processing and subtitle workflow built around OpenAI Whisper.

Requirements: Python, ffmpeg and openai-whisper; model downloads are required separately.

- Single-file and subtitle workflow
- Batch and failure tracking guidance
- Audio quality and human review limits

This is workflow documentation. Prepare the environment, check whisper --help, and process authorized media as described in SKILL.md.

## Install / 安装

~~~bash
unzip audio-transcribe-v1.1.0.zip -d ./skill-install
mkdir -p "$HOME/.agents/skills/audio-transcribe"
cp -R ./skill-install/audio-transcribe/. "$HOME/.agents/skills/audio-transcribe/"
~~~

For Claude Code, replace .agents/skills with .claude/skills. Use the actual installation path in commands. Read SKILL.md and THIRD-PARTY.md before running anything. Dependencies, model weights, accounts and media are not included. No full runtime or all-client compatibility claim is made.

## Package files

- LICENSE.txt: 1115 bytes; SHA-256 8299c7d39816b7abaafd55ce0ab73db564238876116533acfc511ff385ff5c22
- README.md: 2826 bytes; SHA-256 a1e233e5ac694758dc2b1711bcaa88ab304d09f53d12ebb40b661f1f81b18e00
- SKILL.md: 2251 bytes; SHA-256 3fd39ca6627960922205165040d964bff6eb8dcde4032e71b905d25476c6fdeb
- THIRD-PARTY.md: 230 bytes; SHA-256 5043fd1fba191fba1dd6c996ed7236b33f2f1a6933ccc5a2668c2d8db30d5c93