YeJe-cpu/easy-jianying

What it does

Easy JianYing (jyr) is a Python CLI that converts a JSON editing plan into a native JianYing (剪映) or CapCut draft on macOS. The result is an ordinary project with separate video, text, subtitle and audio tracks, transitions, keyframes and beat markers, so you can adjust individual clips and subtitles in the app instead of re-rendering a finished video. It targets editors using Claude Code or Codex-style agents for rough cuts while keeping final control inside JianYing/CapCut.

The tool also provides talking-head rough cutting, local speech recognition for subtitles, music beat detection for rhythm-based cuts, batch generation from templates, and automatic selection of free resources when a value is set to "auto". See the how-it-works documentation for the draft format mechanics.

Quick start

Requirements: a Mac, JianYing Professional (or CapCut), Python 3.10+, and ffmpeg. Push the draft only when JianYing is fully closed.

git clone https://github.com/YeJe-cpu/easy-jianying && cd easy-jianying
pip install -e '.[asr,fonts]'
jyr doctor
./tools/make_fixtures.sh examples/media
jyr build examples/showcase.plan.json --out /tmp/jyr-demo --publish --name "My first AI draft"

The jyr doctor step should report ok; the last command publishes the draft so it appears on the JianYing home page. For CapCut, add --app capcut to doctor, build and inspect commands. The three-minute guide covers the same flow in the original README, including a sample plan that uses second-based durations and Chinese resource names.

To use it from an agent, install the skills under skills/ into Claude Code or Codex and ask for the rough cut to be placed into JianYing.

Version and resource constraints

Drafts are verified against JianYing 11.7.0 and CapCut 9.4.1 on Intel Macs; Apple Silicon is described as theoretically workable but untested, and Windows is untested. JianYing-generated drafts that JianYing has already saved cannot be read back — edit the plan and regenerate, or edit directly in the app. CapCut drafts, by contrast, can be read, and subtitles can be exported as SRT.

By default the tool uses only free built-in resources (175 of 375 cataloged resources). Member transitions, masks and text animations require --member or "member": true. Membership checks are left to JianYing/CapCut; the tool does not bypass them. Not yet implemented: cropping, matting, color grading, curve speed changes, fancy text, and audio effects. Subtitle recognition can contain errors, which can be batch-replaced or handled by your own recognition service.

The project is not an official JianYing tool and is unrelated to ByteDance. The license is PolyForm Noncommercial 1.0.0: personal learning, research and non-commercial use are allowed, but commercial use — including paid courses, paid services or client delivery — is not.

HotRepo's take

The repository has not appeared on GitHub's all-language daily ranking. On the new-project list it recorded 107 stars and 29 forks on 2026-11-02, with a one-hour growth window adding 6 stars and 2 forks. That single window is too short to establish a trend; the useful signal is that the project documents acceptance checks against real JianYing playback and includes a comparison of rule-based rough cutting with manual editing. Those concrete verification notes are more informative than the star count.