AI repository intelligence

FR-AI / ANALYSIS

Why it is worth attention

It bridges AI coding agents with a macOS video tool, enabling complex subtitle workflows entirely through natural language commands for the first time in an open-source package.

Who it is for

AI coding agent users (Claude Code, Codex)
Video content creators and editors
Developers automating subtitle pipelines
Anyone using skills.sh-compatible agents

Use cases

Transcribe and translate spoken content in video files to Chinese or other languages
Add, polish, and review subtitles for talking-head videos
Export translated subtitles in SRT format for distribution
Clean up talking-head footage by removing pauses or filler content

Strengths

Natural language interface eliminates manual CLI learning
Open-source (MIT) and integrates with popular AI agents like Claude Code and Codex
Simple installation via npx or manual git clone
Part of the skills.sh ecosystem, making it easy to discover and link with multiple agents

Considerations

Requires macOS and the BaoCut app (not included) to function
Only works with skills.sh-compatible agents (Claude Code, Codex, etc.)
One-command install needs Node.js, though manual install does not

README quick start

BaoCut agent skill

Give your AI coding agent the power to drive BaoCut — transcribe, add & translate subtitles, review speakers, clean up talking-head video, and export — all from natural language. This is the open-source Agent Skill that wraps the baocut command-line tool.

Works with Claude Code, Codex, and any skills.sh-compatible agent.

Requirements

macOS — BaoCut is a Mac app.
The BaoCut app — install it from baocut.app. The skill drives the CLI bundled inside BaoCut.app; it does not ship the engines.
Node.js — only for the one-command install below (download). The manual steps need no Node.

Install

Recommended — skills.sh

npx skills add JimLiu/baocut -g -a claude-code codex -y

This installs the skill globally to ~/.agents/skills/baocut and links it for Claude Code and Codex.

Manual (no Node.js)

git clone https://github.com/JimLiu/baocut.git ~/.agents/skills/baocut-src
ln -sfn ~/.agents/skills/baocut-src/skills/baocut ~/.agents/skills/baocut
# link it for each agent you use, e.g. Claude Code:
mkdir -p ~/.claude/skills
ln -sfn ~/.agents/skills/baocut ~/.claude/skills/baocut

Use it

Once installed, just ask your agent — for example:

Claude Code — "Transcribe and translate the subtitles of talk.mp4 to Chinese," or the Chinese equivalent 转写并翻译字幕. You can also type /baocut.
Codex — reference the baocut skill in your prompt; it drives the baocut CLI.

Under the hood the agent runs commands like:

baocut --json auto talk.mp4 --lang zh    # transcribe -> polish -> translate
baocut export  --srt --translated --lang zh

If baocut is not on your PATH, use the bundled resolver skills/baocut/bin/baocut (it points at /Applications/BaoCut.app/Contents/MacOS/baocut-cli and tells you to install the app if it is missing).

Layout

skills/baocut/
  SKILL.md          # agent entry point (router)
  references/       # per-task guides (orchestration, editing, export, …)
  bin/baocut        # resolves the CLI inside the installed BaoCut.app

License

MIT.

Description

Open-source Agent Skill that drives the BaoCut macOS app CLI (transcribe · subtitle · translate · cut) from Claude Code, Codex, and other agents

Related repositories

Similar projects matched by category, topics, and programming language.

makecindy

Featured

cindy

Cindy is an open-source AI agent that runs locally on your machine, integrates multiple AI harnesses and models, and provides memory, skills, and automation to perform real work in your projects and apps.

AI & Machine LearningLarge Language Models

958

7-e1even