Complete ComfyUI Workflow Breakdown for Music Videos
Ready_Yam4471 · reddit · 2026-07-07
A creator detailed their first large-scale music video project (transforming "GOLDEN" into a dark universe): using QWEN Edit with multi-angle LoRA to build characters and scenes, then connecting to Wan2.2 for image-to-video, all based on a basic ComfyUI template with a few custom nodes. For 4K upscaling, SeedVR2 yielded poor results (artifacts, ghosting), so they switched to Wan2.2 with Ultimate SD Upscale for better details, though requiring significant VRAM (using RunPod instances). The biggest pain point was QWEN Edit's instability with multiple characters and reference environments, often locking the camera to the reference angle and requiring extensive iteration. The workflow is open-sourced on GitHub.
More from Multimodal
- AIandDesign reveals Runway Act-2 + ElevenLabs voice changer video workflow — AIandDesign · 2026-10-26
- A prompt for generating downhill mountain biking action-sequence AI videos — umesh_ai · 2026-10-26
- Claude scripted, generated and edited an 'emotional AI commercial' end-to-end — CurieuxExplorer · 2026-10-26
- Claude Motion prompt galleries let you copy and recreate trending AI videos — CodeByPoonam · 2026-10-26
- MOSS Image: fully open-source 5B text-to-image model with Apache 2.0 and JSON layout control — Sam Witteveen · 2026-10-26
- Uncanny AI-generated SpongeBob with giant hands floods X — OVolosin82152 · 2026-10-26