What is ComfyUI and why should you learn it in 2026?
ComfyUI is a free, open-source, node-based interface for AI image and video generation. It supports the latest models — Flux, HiDream, WAN 2.2, HunyuanVideo, and more — and ships with first-day support for every new model architecture. With over 114,000 GitHub stars as of 2026, it's the most popular local AI workflow tool in the world.
Unlike paid tools like Midjourney or DALL·E, ComfyUI runs entirely on your own computer — giving you full control, zero subscription fees, and complete privacy. And with the new App Mode and ComfyHub (launched March 2026), you can now turn any workflow into a shareable app that anyone can run without touching a node graph.
This roadmap takes you from installing ComfyUI for the first time all the way to running Flux, creating AI videos, and training your own models — completely free, step by step.
Your ComfyUI Learning Path
Click any level to expand it. Work top-to-bottom for best results. Steps marked 🆕 are new for 2026.
AI Start 1: Beginner Foundations
New to ComfyUI? Start from zero — no experience needed.
AI Boost 2: Text-to-Image Mastery
Level up your text-to-image skills — image manipulation, ControlNet, LoRA, Flux, upscaling, and every major 2026 image model like Qwen Image, Krea 2, and Nano Banana.
AI SkillUp 3: Text-to-Video Mastery
Generate AI video locally with WAN 2.1/2.2, LTXV, HunyuanVideo, and FramePack.
AI Power 4: Advanced Creations
Train your own LoRA and DreamBooth models. Build automations and APIs.