OmniFeed Search

ComfyUI

Node-based app for generating images, video, audio and 3D with open models on your own computer

Visit comfy.org
Screenshot of ComfyUI
ComfyUI, from github.com

Our verdict

#3 of 4 in Local AI image generators · For video, audio and 3D, or building your own pipelines

Pros

  • Desktop app on all three desktop systems, with automatic updates
  • Documents NVIDIA, AMD, Intel and Apple Silicon graphics, plus a processor-only build
  • Lists 3D models, which the other three don't document
  • Saves workflows as files and recovers them from generated images

Cons

  • Comfy Desktop sends anonymous usage data unless you untick it at setup
  • No user accounts documented for sharing one install
  • No canvas with layers described; editing happens through nodes and masks
  • No documented way to download a model from a link you paste

Who should skip:Households where several people need their own accounts on one install, since the docs describe none.

Read the full review →

Rankings and alternatives

Apps for generating images locally scores

Evaluated Nov 7, 2026 Evidence ↓
  • Image model families 5

    5/5 means: Fifteen or more families

    The README's image generation list names 21 model families, from Stable Diffusion 1.5 to Moss Image, and calls the list representative.

  • Graphics cards supported 5

    5/5 means: NVIDIA, AMD, Apple Silicon and Intel Arc, plus a documented processor-only mode

    The manual install supports NVIDIA, AMD, Intel and Apple Silicon GPUs (plus Ascend NPUs), and the Windows portable build runs on the processor alone.

  • Setup effort 5

    5/5 means: A graphical installer or desktop app sets everything up and installs updates from inside the app

    Comfy Desktop installs on Windows, macOS (Apple Silicon) and Linux, the README recommends it to new users, and it installs its own updates by default.

  • Editing tools 3

    3/5 means: Three of the five

    Built-in tools for inpainting, outpainting and upscaling are documented. Masks and compositing are listed, but no canvas with layers or regional prompting is described, so three of the five count.

  • Custom workflows 5

    5/5 means: As 4, plus running a saved workflow from a simple form or the main screen without the graph

    The whole app is a node graph; workflows save and load as JSON and can be recovered from generated images, and App Mode exposes a workflow through a simple UI.

  • Add-ons 5

    5/5 means: As 4, plus updating installed add-ons from inside the app

    ComfyUI-Manager searches, installs, updates, removes and disables custom node packs from inside the app. On a manual install it is turned on with a launch flag.

  • Getting models 2

    2/5 means: The app downloads only models from its own list, such as starter models or the ones a template needs

    Comfy Desktop downloads the models a template or a loaded workflow needs, and ComfyUI-Manager downloads models from its own lists. The docs don't describe downloading any model from a link you paste.

  • What leaves your computer 2

    2/5 means: Generates locally, but the recommended install sends usage data by default

    The recommended install, Comfy Desktop, shares anonymous usage data unless you untick the box at setup; it never collects prompts or images. Paid partner nodes for closed models are optional, and the --offline flag keeps everything local on other installs.

  • Video and audio 5

    5/5 means: Video, audio and 3D

    The README lists local video models (Wan, LTX-Video, HunyuanVideo and more), audio models (ACE-Step, Stable Audio) and 3D models (Hunyuan3D, TripoSplat).

  • Several users 1

    1/5 means: No sign-in or user accounts documented

    The README's feature list and the Comfy Desktop docs we read describe no user accounts or sign-in for a shared local install.

    Rests partly on what the sources don't mention, which is a weaker signal than what they say.

Evidence and sources

Apps for generating images locally scores

Image model families 1 quote
  • “Image generation: Stable Diffusion 1.5, SDXL, SD3.5, Flux.1, Flux.2, Qwen Image, Z-Image, Hunyuan Image 2.1, HiDream, Lumina Image 2.0, Chroma, Anima, LongCat Image, Ideogram 4, Krea 2, MageFlow, Microsoft Lens, PixelDiT, Kandinsky 5, Ernie Image and Moss Image.”
    github.com/Comfy-Org/ComfyUI · checked
Graphics cards supported 3 quotes
  • “Supports all operating systems and GPU types (NVIDIA, AMD, Intel, Apple Silicon, Ascend).”
    github.com/Comfy-Org/ComfyUI · checked
  • “There is a portable standalone build for Windows that should work for running on Nvidia GPUs or for running on your CPU only.”
    github.com/Comfy-Org/ComfyUI · checked
  • “Intel Arc GPU users can install native PyTorch with torch.xpu support using pip.”
    github.com/Comfy-Org/ComfyUI · checked
Setup effort 3 quotes
Editing tools 1 quote
  • “Built-in tools for inpainting, outpainting, reference conditioning, masks and compositing, model merging, upscaling, frame interpolation, segmentation, depth estimation, and media processing.”
    github.com/Comfy-Org/ComfyUI · checked
Custom workflows 3 quotes
  • “A visual node graph for building and reusing image, video, audio, 3D, and text workflows without code.”
    github.com/Comfy-Org/ComfyUI · checked
  • “Save and load workflows as JSON, or recover complete workflows and seeds from supported generated media.”
    github.com/Comfy-Org/ComfyUI · checked
  • “The most sophisticated workflows can be exposed through a simple UI thanks to App Mode.”
    github.com/Comfy-Org/ComfyUI · checked
Add-ons 3 quotes
Getting models 3 quotes
What leaves your computer 4 quotes
  • “Help improve Comfy by sharing anonymous usage data: optional and ticked by default. Comfy Desktop sends anonymous usage data and never collects workflow contents, prompts, generated media, or personal information.”
    docs.comfy.org/installation/desktop/overview · checked
  • “Runs fully offline: core does not download anything unless you request it. Use --offline to disable the optional paid Comfy API nodes and force all built-in functionality to stay offline, or --disable-partner-nodes to only disable the API nodes.”
    github.com/Comfy-Org/ComfyUI · checked
  • “Partner nodes provide access to the best closed source models such as Nano Banana, Seedance, Hunyuan3D, etc.”
    github.com/Comfy-Org/ComfyUI · checked
  • “The desktop app is the easiest and best way to use ComfyUI for new users.”
    github.com/Comfy-Org/ComfyUI · checked
Video and audio 3 quotes
  • “Video generation: Wan 2.1 and 2.2, LTX-Video 2 and 2.3, HunyuanVideo 1.5, Kandinsky 5 Video, CogVideoX, Cosmos Predict2, Bernini-R, SCAIL 2, and Mochi.”
    github.com/Comfy-Org/ComfyUI · checked
  • “Audio generation: ACE-Step 1.5, Stable Audio 3, MiniMax Music 3 and Yue 2.”
    github.com/Comfy-Org/ComfyUI · checked
  • “3D and vision: Hunyuan3D 2.1, TripoSplat, SeedVR2, SUPIR, Depth Anything 3, MoGe, SAM 3 and 3.1, RT-DETRv4, and BiRefNet.”
    github.com/Comfy-Org/ComfyUI · checked
Several users 2 quotes
  • “A visual node graph for building and reusing image, video, audio, 3D, and text workflows without code.”
    github.com/Comfy-Org/ComfyUI · checked
  • “The desktop app is the easiest and best way to use ComfyUI for new users.”
    github.com/Comfy-Org/ComfyUI · checked

Facts

License 1 quote
Runs on 1 quote
Needs 2 quotes
  • “There is a portable standalone build for Windows that should work for running on Nvidia GPUs or for running on your CPU only.”
    github.com/Comfy-Org/ComfyUI · checked
  • “Can run even the biggest open source models on as low as 4GB vram + 8GB ram relatively quickly (saturating your GPU compute) using our state of the art asynchronous weight streaming technology.”
    github.com/Comfy-Org/ComfyUI · checked

Notability

Company value 1 quote

Other quotes

Not cited by a score or fact 4 quotes