nateherkai/herk-edit-kit

What it solves

This project provides an automated workflow for video editing, specifically designed to handle repetitive tasks like cutting dead air, identifying mistakes, syncing clips to music beats, and generating motion graphics. It enables users to offload the technical labor of video production to AI agents within coding environments.

How it works

The kit functions as a set of specialized skills for AI assistants like Codex and Claude Code. It leverages the HyperFrames framework to execute editing commands, such as cut-silences, video-storytelling, or short-form-edit. Users provide raw footage and instructions, and the agent uses these predefined skills to perform tasks like transcript retiming, captioning, and applying motion-graphic templates built with GSAP. It supports optional integration with external services for transcription and asset generation, but can also run locally.

Who it’s for

It is intended for content creators and video editors who use AI-assisted coding tools and want to automate their editing pipeline. It is particularly useful for those producing short-form content, YouTube videos, or motion-design showreels who want to maintain a consistent stylistic "house" look.

Highlights

• Includes 15 specialized editing skills for tasks like beat-syncing, mistake detection, and captioning.
• Provides over 400 motion-graphics templates and CSS-based design tokens.
• Supports agent-guided workflows for both long-form storytelling and short-form reels.
• Includes a comprehensive editing playbook that codifies stylistic rules into actionable agent instructions.
• Offers a synthetic starter project that allows users to test the editing pipeline without needing their own footage or API keys.