{"id":32496,"date":"2025-04-06T16:05:33","date_gmt":"2025-04-06T08:05:33","guid":{"rendered":"https:\/\/www.1ai.net\/?p=32496"},"modified":"2025-04-06T16:10:06","modified_gmt":"2025-04-06T08:10:06","slug":"comfyui%e9%83%a8%e7%bd%b2%e6%95%99%e7%a8%8b%ef%bc%8ccomfyui%e5%ae%89%e8%a3%85%e4%bf%9d%e5%a7%86%e7%ba%a7%e6%8c%87%e5%8d%97","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/32496.html","title":{"rendered":"ComfyUI Deployment Tutorial, ComfyUI Installation Nanny Guide"},"content":{"rendered":"

If you're still drawing with the traditional Stable Diffusion WebUI, you're really out of the game! Nowadays, the hottest tool in the AI mapping community - theComfyUI<\/a><\/strong>The game has been quietly \"out of the loop\" from the hands of tech geeks and has become a dream bag for advanced players. This article is tailor-made for you!ComfyUI Installation Nanny Tutorial<\/strong>, from the little white guy to the veteran driver, take it all!<\/p>\n

I. What is ComfyUI?<\/h2>\n

ComfyUI, that's its full name, is on GitHub at https:\/\/github.com\/comfyanonymous\/ComfyUI.<\/p>\n

\"ad888a78j00suaeaf004td000u000d5m\"<\/section>\n

To put it plainly, it is a node-based workflow AI image generation tool, specifically tailored for diffusion models like Stable Diffusion. You can drag and drop like Lego, connect different nodes, such as model loading, cue input, sampler, etc., to build your own \"image factory\".<\/p>\n

\"4821b639j00suaeaf0023d000n700f5m\"<\/section>\n

Compared to the traditional way of operation, the most important feature of ComfyUI is - theFlexible and controllable<\/strong>You can control every detail of the parameters. You have precise control over every detail parameter and complete autonomy in building your workflow.That's what a master should do!<\/strong><\/p>\n

Core highlights at a glance:<\/h3>\n
    \n
  • \n

    nodal operation<\/strong>: It's like doing programming like building blocks, with modular control every step of the way;<\/p>\n<\/li>\n

  • \n

    low graphics memory threshold<\/strong>The graphics card 8GB can also take the fly, who said AI drawing can only rely on \"home mine\";<\/p>\n<\/li>\n

  • \n

    Batch processing capability<\/strong>: Process automation, batch image\/video generation made easy;<\/p>\n<\/li>\n

  • \n

    Highly scalable<\/strong>: The plugin ecosystem is growing rapidly, asControlNet<\/span>, AnimateDiff, and a whole lot more.<\/p>\n<\/li>\n<\/ul>\n

    II. ComfyUI VS Stable Diffusion WebUI<\/h2>\n

    Many of you will ask, so I'm already using Stable Diffusion WebUI, do I still need ComfyUI?<\/p>\n

    Don't worry, you'll know the answer after reading this comparison table:<\/p>\n\n\n\n\n\n\n\n\n\n\n
    \n
    comparison dimension<\/section>\n<\/th>\n
    		\n
    ComfyUI<\/section>\n<\/th>\n
    		\n
    Stable Diffusion WebUI<\/section>\n<\/th>\n<\/tr>\n<\/thead>\n
    interface style<\/strong><\/td>\n\n
    Nodal Visualization Workflow<\/section>\n<\/td>\n
    		\n
    Traditional Web Form Buttons<\/section>\n<\/td>\n<\/tr>\n
    learning threshold<\/strong><\/td>\n\n
    Slightly higher, need to understand node logic<\/section>\n<\/td>\n
    		\n
    Low, newbie friendly<\/section>\n<\/td>\n<\/tr>\n
    hardware requirement<\/strong><\/td>\n\n
    Lower video memory requirements (\u22658GB)<\/section>\n<\/td>\n
    		\n
    Recommended 12GB+ video memory<\/section>\n<\/td>\n<\/tr>\n
    functional freedom<\/strong><\/td>\n\n
    Highly customizable, unlimited combinations<\/section>\n<\/td>\n
    		\n
    Fixed and slightly rigid functions<\/section>\n<\/td>\n<\/tr>\n
    plug-in ecology<\/strong><\/td>\n\n
    Developing with high potential<\/section>\n<\/td>\n
    		\n
    Mature, but slow to innovate<\/section>\n<\/td>\n<\/tr>\n
    Suitable for the scene<\/strong><\/td>\n\n
    Batch processing, process studies<\/section>\n<\/td>\n
    		\n
    Rapid single-image creation<\/section>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    One sentence summary:Beginner's Chart One Fun, Advanced Chart ComfyUI<\/strong>.<\/p>\n

    Third, the pre-installation preparation: hardware + system + drive<\/h2>\n

    The hardware requirements are not really high if you want to play with ComfyUI:<\/p>\n

    \u2705 Local Deployment Recommended Configuration:<\/h3>\n
      \n
    • \n

      Graphics card: NVIDIA \u2265 1080ti (\u2265 8GB video memory); AMD users need to goROCm<\/span>Programs;<\/p>\n<\/li>\n

    • \n

      Storage: 40GB minimum, 100GB or more recommended (SSD accelerated experience is better);<\/p>\n<\/li>\n

    • \n

      System:Windows 10<\/span>\/11, compatible and stable without pit.<\/p>\n<\/li>\n<\/ul>\n

      \u2705 Software dependencies:<\/h3>\n
        \n
      • \n

        Graphics Card Driver<\/strong>: NVIDIA needs to support CUDA 11.8+ without installing CUDA Toolkit alone;<\/p>\n<\/li>\n

      • \n

        VS Build Tools<\/strong>: Essential for Windows platform installations;<\/p>\n<\/li>\n

      • \n

        AMD users<\/strong>: WSL 2 + ROCm needs to be configured.<\/p>\n<\/li>\n<\/ul>\n

        Fourth, ComfyUI mainstream installation method full analysis<\/h2>\n

        \ud83d\udfe2 Way 1: Desktop ComfyUI shell (recommended for newbies, future trend)<\/h3>\n
        \n

        Current version: ComfyUI Setup 0.4.32
        Download: https:\/\/www.comfy.org\/download<\/p>\n<\/blockquote>\n

        Here are the steps:<\/h4>\n
          \n
        1. \n

          (Optional) DNS acceleration<\/strong>modifications\u00a0C:\\Windows\\System32\\drivers\\etc\\hosts<\/code>, add:<\/p>\n

          Copy edit 185.199.110.133 raw.githubusercontent.com<\/pre>\n<\/li>\n
        2. \n
          Download the installation package<\/strong>double-click to install\u00a0ComfyUI Setup 0.4.32 - x64.exe<\/code>The program will enter the graphical interface and start the automatic installation.<\/p>\n<\/li>\n
        3. \n
          Just default all the way.<\/strong><\/p>\n
            \n
          • \n
            GPU selection: do not check CPU;<\/p>\n<\/li>\n
          • \n
            Installation path: Default is fine;<\/p>\n<\/li>\n
          • \n
            Mirror settings: Replace with domestic source to speed up download.<\/p>\n
            arduino copy edit Python mirror: https:\/\/gitcode.com\/gh_mirrors\/py\/python-build-standalone\/releases\/PyTorch\u955c\u50cf: https:\/\/mirrors.tuna.tsinghua.edu .cn\/anaconda\/cloud\/pytorch\/<\/pre>\n<\/li>\n<\/ul>\n<\/li>\n
          • \n
            Click Install and wait for auto-completion<\/strong><\/p>\n
            \"e10e4579j00suaeae000sd000cv004km\"<\/section>\n<\/li>\n
          • \n
            configuration model<\/strong>When the default workflow is run, it prompts for missing models such asv1-5-pruned-emaonly.ckpt<\/code>(SD1.5 large model).<\/p>\n
            Download the model and put it in:<\/p>\n
            Makefile copy edit D:\\AI Files\\ComfyUI\\models\\checkpoints\\<\/pre>\n
            Then restart ComfyUI and re-run the workflow, theSee the image generation page and you're done!<\/strong><\/p>\n<\/li>\n<\/ol>\n
            \"f9e2dc60j00suaeaf002md000o800gmm\"<\/section>\n
            \"050f5080j00suaeaf004bd000o300gom\"<\/section>\n
            \ud83d\udfe1 Way 2: Windows Portable Standalone Integration Pack (current mainstream play)<\/h3>\n
            \n
            Download: https:\/\/github.com\/comfyanonymous\/ComfyUI\/releases\/latest\/download\/ComfyUI_windows_portable_nvidia.7z<\/p>\n<\/blockquote>\n
            Here are the steps:<\/h4>\n
              \n
            1. \n
              Unzip the file.Note that the path cannot contain Chinese characters<\/strong>.<\/p>\n<\/li>\n
            2. \n
              users who already have a WebUI model library.directly reusable<\/strong>The download does not need to be repeated:<\/p>\n
                \n
              • \n
                Modify the configuration:<\/p>\n
                yaml copy edit rename extra_model_paths.yaml.example to extra_model_paths.yaml edit as follows (replace path with your actual local path): a111: base_path: D:\\AI Files\\sd-webui-aki-v4.10\\ controlnet: D:\\AI Files\\sd-webui-aki-v4.10\\extensions\\sd-webui-controlnet\\models<\/pre>\n
                \"174314adj00suaeaf003qd000o500btm\"<\/section>\n
                <\/pre>\n<\/li>\n<\/ul>\n<\/li>\n
              • \n
                double click\u00a0run_nvidia_gpu.bat<\/code>\u00a0Start ComfyUI.<\/p>\n<\/li>\n
              • \n
                Install the management plug-in artifact [ComfyUI-Manager]:<\/p>\n
                bash copy edit into custom_nodes folder \u2192 Shift + Right click \u2192 open PowerShell and run command: git clone https:\/\/github.com\/ltdrdata\/ComfyUI-Manager.git<\/pre>\n<\/li>\n
              • \n
                Restart ComfyUI, done!<\/p>\n<\/li>\n<\/ol>\n
                \"09d4eb09j00suaeaf003od000oe00f2m\"<\/section>\n
                \ud83d\udd35 Way 3: Eisai Integration Pack (best for the fear of trouble party)<\/h3>\n
                \n
                Integration pack release page: https:\/\/www.bilibili.com\/video\/BV1Ew411776J\/
                Current version: ComfyUI-aki-v1.6.7z (with decryption code)<\/p>\n<\/blockquote>\n
                \"cb8021f1j00suaeaf00ald000og00enm\"<\/section>\n
                Installation Steps:<\/h4>\n
                  \n
                1. \n
                  Unzip to a Chinese-free path.<\/p>\n<\/li>\n
                2. \n
                  Copy:<\/p>\n
                    \n
                  • \n
                    Model files to the\u00a0ComfyUI\\models\\checkpoints\\<\/code><\/p>\n<\/li>\n
                  • \n
                    ControlNet model to\u00a0ComfyUI\\models\\controlnet\\<\/code><\/p>\n<\/li>\n<\/ul>\n<\/li>\n
                  • \n
                    double click\u00a0Eisai Launcher.exe<\/code>\u00a0\u2192 Click on [One Touch Start]<\/p>\n<\/li>\n
                  • \n
                    automatically open the browser interface.Start creating straight away<\/strong>!<\/p>\n<\/li>\n<\/ol>\n
                    V. Commonly used catalog descriptions (suggested collections)<\/h2>\n\n\n\n\n\n\n\n\n\n\n\n\n
                    \n
                    catalogs<\/section>\n<\/th>\n
                    		\n
                    corresponds English -ity, -ism, -ization<\/section>\n<\/th>\n<\/tr>\n<\/thead>\n
                    models\/checkpoints\/<\/code><\/td>\n\n
                    Master models go here, such as SD1.5, SDXL<\/section>\n<\/td>\n<\/tr>\n
                    models\/loras\/<\/code><\/td>\n\n
                    LoRA fine-tuning model<\/section>\n<\/td>\n<\/tr>\n
                    models\/controlnet\/<\/code><\/td>\n\n
                    ControlNet related models<\/section>\n<\/td>\n<\/tr>\n
                    models\/vae\/<\/code><\/td>\n\n
                    VAE model for color, detail control<\/section>\n<\/td>\n<\/tr>\n
                    embeddings\/<\/code><\/td>\n\n
                    Text Embedding Model<\/section>\n<\/td>\n<\/tr>\n
                    custom_nodes\/<\/code><\/td>\n\n
                    Plugins go here (e.g. AnimateDiff, etc.)<\/section>\n<\/td>\n<\/tr>\n
                    output\/<\/code><\/td>\n\n
                    Generated images\/videos are automatically saved here<\/section>\n<\/td>\n<\/tr>\n
                    input\/<\/code><\/td>\n\n
                    Enter a reference chart or cue word material<\/section>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
                    ComfyUI is the future and not to be missed!<\/h2>\n
                    ComfyUI is not just a tool, it's a complete AI image workflow framework. It is not only suitable for players who want to deeply customize the generation process, but also is becoming more and more \"approachable\". As the plug-in ecosystem continues to grow, the desktop shell is becoming more and more perfect.ComfyUI is taking a step towards mainstreaming.<\/strong><\/p>\n
                    To run on the AI mapping track, now is the best time to get on board!<\/p>\n
                    Go ahead and try it, once you get started, you simply can't stop!<\/strong><\/p>","protected":false},"excerpt":{"rendered":"
                    If you are still using the traditional Stable Diffusion WebUI drawing, then you are really OUT! Now the hottest tool in the AI drawing circle - ComfyUI, has quietly from the hands of technical geeks \"out of the circle\", and become the dream of advanced players in the love package. This article is tailored for you to install ComfyUI nanny-level tutorials, from small white to old drivers, through to take! First, what is ComfyUI? ComfyUI, the full name is called this, its GitHub address is: https:\/\/github.com\/comfyanonymous\/ComfyUI. To put it bluntly, it is a node-based workflow AI image generation tool, specialized for Stable D<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[1989,4749,2894,2658],"collection":[],"class_list":{"0":"post-32496","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-jiaocheng","7":"category-baike","8":"tag-comfyui","10":"tag-2894","11":"tag-2658"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/32496","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=32496"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/32496\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=32496"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=32496"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=32496"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=32496"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}