Brethof AI Reviews

ComfyUI review

ComfyUI logo

A game changer for AI image, video and audio

Not sponsoredDesign & Graphics · Video & Audio

Tested on Linux · Free · Open source

ComfyUI is a free, open-source app for AI images, video, audio and 3D. It runs on your own graphics card or through cloud models. We use it all the time; parts of this video are made in it.

The good

  • Templates for everything. 602 ready-made workflows for image, video, audio and 3D. Pick one and get results without touching the graph.
  • Runs on any PC. Small quantized models work even on modest graphics cards, and without a GPU, 357 templates use partner cloud models, paid per run.
  • New models first. New AI models show up in ComfyUI fast, often the day they're released, ready as templates.
  • App Mode. App Mode turns a finished workflow into a simple form anyone can use, with no graph in sight.

The not so good

  • Local needs a GPU. Running ComfyUI locally needs a graphics card, and the newest full-size models need a strong one. Through the cloud, any PC works.
  • Learning never ends. Beyond the templates you learn the graph, and every workflow and node pack is different. Custom nodes keep changing, so the learning never really ends.
  • Third-party custom nodes. Custom nodes are code written by strangers, and malware has already shipped through the official node registry once. Only install what you trust.

Our verdict

Strongly recommended

It's a game changer that opens state-of-the-art AI generation to anyone. Templates give results with no learning, the graph is there when you want more, and it runs on modest hardware or in the cloud; quality scales with your setup.

Reviewed: ComfyUI 0.40.0 · facts checked 2026-10-14

← All reviews

Hear it when it ships

New releases, real benchmarks and the occasional deep-dive. No spam, unsubscribe in one click.

Everything we build

External:   YouTube · GitHub