How whatdreamscost/whatdreamscost-comfyui works
This repository provides custom ComfyUI nodes and browser interfaces for preparing media and building LTX video-generation timelines. The central workflow is LTX Director: users edit video, image, prompt, and audio segments, then use the node with ComfyUI generation; the guide node adds image and motion guidance. Separate nodes load and trim audio/video, insert keyframes, sequence image guides, and estimate spoken-dialogue duration. The excerpts show Director references to prompt-relay and model-patching code; browser-to-backend wiring and several unsampled node internals are not shown, so those connections are omitted.
Main entry points: __init__.py ComfyUI Nodes · ltx_director.py Timeline Generation · ltx_director_guide.py Guide Processing · patches.py Attention Patches
Components 20
LTX Director
- Timeline Generation
ltx_director.py - Timeline Editor
js/ltx_director.js - Prompt Relay
prompt_relay.py - Attention Patches
patches.py - Guide Processing
ltx_director_guide.py - Guide Interface
js/ltx_director_guide.js
Media Tools
- Audio Loader
load_audio_ui.py - Audio Interface
js/load_audio_ui.js - Video Loader
load_video_ui.py - Video Interface
js/load_video_ui.js
Image Guidance
- Multi Image Loader
multi_image_loader.py - Image Interface
js/multi_image_loader.js - LTX Keyframer
ltx_keyframer.py - Keyframer Interface
js/ltx_keyframer.js - LTX Sequencer
ltx_sequencer.py - Sequencer Interface
js/ltx_sequencer.js
Workflow Utilities
- Speech Length
speech_length_calculator.py - Speech Interface
js/speech_length_calculator.js
Other components
- Creator
- ComfyUI Nodes
__init__.py
Connections 12
- Creator → ComfyUI Nodes: uses nodes
README.md - ComfyUI Nodes → Timeline Generation: registers
__init__.py - ComfyUI Nodes → Guide Processing: registers
__init__.py - ComfyUI Nodes → Audio Loader: registers
__init__.py - ComfyUI Nodes → Video Loader: registers
__init__.py - ComfyUI Nodes → Multi Image Loader: registers
__init__.py - ComfyUI Nodes → LTX Keyframer: registers
__init__.py - ComfyUI Nodes → LTX Sequencer: registers
__init__.py - ComfyUI Nodes → Speech Length: registers
__init__.py - Timeline Generation → Prompt Relay: references
ltx_director.py - Timeline Generation → Attention Patches: references
ltx_director.py - Guide Processing → Attention Patches: references
ltx_director_guide.py