WhatDreamsCost/WhatDreamsCost-ComfyUI — reverse-engineered prompt
Reverse engineered prompt
Build me a ComfyUI custom node pack centered around an LTX Director style timeline editor for AI video creation and editing. I want it to feel like a mini video editor inside ComfyUI, where I can arrange shots on a timeline, use prompts and keyframes to control what happens, support first middle and last frames, and work with video audio and images in one place.
It should also include the extra helper nodes this project suggests, like an easy video loader with preview trim and crop tools, an audio loader with simple trimming, a speech length calculator for estimating video duration from dialogue, prompt relay support, and a multi image loader. The timeline should let me split combine extend and retake sections of a video, blend imported audio with generated audio, and save or load the whole timeline as a json file.
Please include example workflows and make the UI polished and beginner friendly. If you need details, look up the current ComfyUI and LTX video docs online.
Want more depth? Deep Reverse