You can create videos from images or text descriptions using ComfyUI and three different engines. Each engine is built for a specific task. Wan 2.2 gives the best film-level quality. FramePack makes very long videos even on low VRAM. AnimateDiff is fast and lets you control camera motion.
This tool helps anyone who needs video content from a simple character image or a text idea. It picks the right engine based on your needs and your computer's power. You can make talking heads, motion animations, or smooth transitions between two frames.
Global
mkdir -p ~/.claude/skills/comfyui-video-pipelineProject
mkdir -p .claude/skills/comfyui-video-pipelineSource Repository
Remotion Best Practicesremotion-dev/skills
Best practices for building videos with React and Remotion
Remotion Renderhalt-catch-fire/skills
Render React code to MP4 videos with Remotion and the belt CLI
Ai Video Generationhalt-catch-fire/skills
Create AI videos from text or images with 40+ models via inference.sh CLI
Ai Image Generationhalt-catch-fire/skills
Create AI images from text with 50+ models via simple CLI commands
Image To Videoagentspace-so/runcomfy-agent-skills
Pick the right AI model to animate your image into video
Nano Banana 2agentspace-so/runcomfy-agent-skills
Create rapid image drafts with strong typography using Google Nano Banana 2 on RunComfy
Nano Banana Editagentspace-so/runcomfy-agent-skills
Edit images with AI preserve subjects swap backgrounds batch up to 20
Flux Kontextagentspace-so/runcomfy-agent-skills
Flux Kontext Pro edits one image part while keeping the rest identical