Local AI Music Videos in ComfyUI - No Custom Nodes, No Cloud

Learn to create complete AI music videos locally on your computer using only ComfyUI's built-in templates without requiring custom nodes, cloud services, or paid subscriptions. Master the three-step process of generating character portraits with Flux image templates, creating original music from text prompts using the Text-to-Song audio template, and animating images to music with the WAN 2.2 sound-to-video workflow. Discover why close-up portraits work better than full-body shots for lip-sync animation, understand the practical 60-second song length limits for optimal quality, and learn essential techniques for managing VRAM, unloading models between workflows, and using descriptions with negative prompts to reduce motion artifacts and broken limbs. Explore chunk length and batch size settings for longer clips, get hands-on experience with ComfyUI 0.3.68 workflows that remain compatible with newer builds, and understand the GPU requirements and model management needed for successful local AI video generation.

Syllabus

00:00 Local AI music video overview
00:34 ComfyUI version and full workflow plan
01:16 Creating the actor image with Flux templates
02:32 Mobile option – Rythmix AI music app sponsor segment
04:42 Testing close-up vs full-body portraits in ComfyUI
05:06 Generating music with the Text-to-Song audio template
06:31 Best song length and quality limits around 60 seconds
06:38 Building the WAN 2.2 sound-to-video workflow
07:36 Loading audio, loading the portrait, and basic setup
08:44 Managing models, VRAM, and using ComfyUI Manager
09:34 Descriptions, prompts, and negative prompts for cleaner motion
11:21 Chunk length, batch size, and rendering the full clip
12:09 Final result and wrap-up