Deploy ComfyUI on Alibaba Cloud Function Compute for Low‑Cost AI Image Generation
This guide explains how to quickly deploy the node‑based AI image generation tool ComfyUI on Alibaba Cloud Function Compute, covering prerequisites, step‑by‑step configuration, NAS integration, workflow loading, performance tips, and resource cleanup to achieve efficient, serverless AI drawing.
Background
ComfyUI is a node‑based UI for Stable Diffusion that enables highly customizable image‑generation workflows. It is used for e‑commerce posters, anime‑to‑real conversion, photo restoration, portrait customization, stylized filters, and video generation, but requires a powerful GPU for local deployment.
Why Deploy on Alibaba Cloud Function Compute (FC)
Deploying ComfyUI on FC provides scalable GPU resources, reduces hardware costs, and simplifies setup through a serverless, pay‑as‑you‑go model.
One‑Click Deployment Procedure
Open the Function Compute console and go to the Application Center (FC 3.0).
Create a new application under the Artificial Intelligence category, selecting the template “Workflow AI Image Generation – ComfyUI”.
Select a region (Hangzhou or Shanghai) and keep the default options.
Confirm and deploy ; the environment is provisioned in about one minute.
After deployment, click the provided WebUI domain to access ComfyUI. Keep the domain private to avoid unexpected charges.
Initial Workflow Test
Open the WebUI, wait up to 10 seconds if the page initially fails to load, then click Queue Prompt . The first image generation may take ~30 seconds due to a cold start; subsequent generations typically finish in 2‑5 seconds.
Loading the “Clay Filter” Workflow
Download the preset clay workflow.zip from
https://labfileapp.oss-cn-hangzhou.aliyuncs.com/函数计算/基于函数计算部署ComfyUI绘图平台/clay%20workflow.zipand unzip.
In ComfyUI, click Load and import the JSON file.
Upload the source image via Load Image → Choose File .
Press Queue Prompt to generate the stylized output. Portrait images give the best results.
NAS Integration for Model Persistence
Mount a File Storage NAS to persist models and workflow files. Set the local directory to /mnt/auto or /mnt/auto/comfyui. Use a high‑performance NAS instance for large model I/O.
Installing Custom Nodes
Use the ComfyUI‑Manager to install additional nodes, e.g., the Chinese translation plugin AIGODLIKE-COMFYUI-TRANSLATION. Search for the node name, click Install , then Restart and refresh the page.
Accelerating Python Dependency Installation
Edit /mnt/auto/comfyui/root/.pip/pip.conf to use Alibaba Cloud’s PyPI mirror:
[global]
index-url = http://mirrors.aliyun.com/pypi/simple/
[install]
trusted-host = https://mirrors.aliyun.comHandling Missing Nodes
If a workflow references unavailable nodes, install them via the manager or refer to the ComfyUI Guides (https://comfyui-guides.runcomfy.com/) for troubleshooting.
ControlNet Example
The guide includes a ControlNet workflow JSON and model download instructions to demonstrate the visual difference between using ControlNet and direct generation.
Resource Cleanup
When ComfyUI is no longer needed, delete the Function Compute application (no further charges) and detach or delete the NAS file system (storage fees persist while data remains).
References
Function Compute console: https://account.aliyun.com/login/login.htm?oauth_callback=https%3A%2F%2Ffcnext.console.aliyun.com%2Foverview
ComfyUI‑Translation plugin repository: https://github.com/AIGODLIKE/AIGODLIKE-COMFYUI-TRANSLATION
ComfyUI Guides: https://comfyui-guides.runcomfy.com/
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Alibaba Cloud Native
We publish cloud-native tech news, curate in-depth content, host regular events and live streams, and share Alibaba product and user case studies. Join us to explore and share the cloud-native insights you need.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
