Skip to main content
While you can run ComfyUI on individual machines, teams in production will typically host ComfyUI on a cluster of GPUs to maximize production scale. ComfyUI offers two deployment methods: the Developer Platform gives you a fast, managed setup on infrastructure managed by ComfyUI, and self-hosted ComfyUI lets you run ComfyUI on your own machine or infrastructure. Comfy Cloud also has an API that you can use with minimal setup.

Serverless API

Deploy a versioned ComfyUI environment as a managed endpoint with on-demand GPU capacity.

Self-Host

Run ComfyUI on your own hardware, a rented GPU, or a container. Full control.

​
Which deployment should I use?

​
Developer Platform (Serverless API)

The Developer Platform is the managed option. Choose it when you want to:
  • Start fast with minimal setup
  • Reduce infrastructure and maintenance work
  • Still have deep control over which custom nodes, models, and Python dependencies are deployed
  • Configure how you want to scale, from 0 to 1000s of GPUs
  • Use management tools such as log visibility and configuration tools

Serverless API

Build a versioned ComfyUI environment and deploy it as a managed endpoint.

​
Self-Hosted ComfyUI

Self-hosting gives you control over how ComfyUI runs and how you manage your environment. Choose it when you want to:
  • Run ComfyUI on your own infrastructure
  • Finely control upgrades, configuration, and security
  • Design for custom scaling or platform requirements

Self-Hosting Options

Run ComfyUI on your own hardware, a GPU cloud, or a container.
Both methods run the same workflows. A workflow you build and test on a self-hosted instance can move to the Developer Platform without changes, because every surface accepts the same API format.

​
Next step: run workflows against it

Once a deployment is live, see Running Workflows to call it from your application. The Comfy SDKs work against every deployment target; self-hosted instances additionally need the API Proxy during the v2 beta.