Headless Jianying: native drafts from Python, no SDK

What it does

This is a Python bridge that writes native Jianying Pro (macOS) draft files from structured JSON editing plans, registers them locally so the app opens them as genuine projects, and can spawn isolated MP4 exports through Jianying’s own engine. It also creates independent copies of existing drafts for automated modification without overwriting originals. The project is explicitly not an official SDK, not a standalone editor, and ships none of the official program libraries, account data, or effect resources.

The interesting bit

Rather than rendering video itself, the tool writes Jianying’s internal binary draft format directly—producing 23-track timelines with more than 150 editable segments that survive round-trips through the native UI. The build pipeline even checksums the official app’s libraries and enforces exact version matches (11.5.0 / 11.4.2) to avoid subtle format drift.

Key highlights

• Generates native drafts with editable text, multi-track video, PIP, subtitles, speed changes, and local audio—materials remain editable after handoff.
• Exports H.264/AAC MP4 through an isolated, offline-native process that reads no account data and touches only the local engine.
• Modifies existing drafts by validating sources and spinning up independent copies, never overwriting the original project.
• Verifies official app signatures, library hashes, and version constants before running; unsupported Jianying builds are rejected outright.
• Ships as a separable “Agent Skill” with its own entry point, though it still needs the core engine checkout and a locally installed Jianying Pro.

Caveats

• Requires Apple Silicon, macOS 26.0+, and an exact Jianying Pro version match; the preview has not yet passed a clean-machine install test.
• Some native effects, curved animations, and compound clips are partially supported or experimental, so “editable” does not mean “pixel-perfect” to the original vision.
• The license is personal/non-commercial only; commercial use requires explicit author permission and the project is not MIT/Apache-2.0.

Verdict

Worth a look if you are building an AI content pipeline that needs to hand off structured timelines to a human-friendly NLE for finishing. Skip it if you need a cross-platform, fully open-source video renderer, or if your Jianying install is not meticulously version-pinned.