ComfyUI docker images for use in GPU cloud and local environments. Includes AI-Dock base for authentication and improved user experience.
Shell
1,067
111 commits
updated Nov 4, 2024
Run ComfyUI in a highly-configurable, cloud-first AI-Dock container.
[!NOTE] These images do not bundle models or third-party configurations. You should use a provisioning script to automatically configure your container. You can find examples, including
SD3&FLUX.1setup, inconfig/provisioning.
All AI-Dock containers share a common base which is designed to make running on cloud services such as vast.ai as straightforward and user friendly as possible.
Common features and options are documented in the base wiki but any additional features unique to this image will be detailed below.
The :latest tag points to :latest-cuda and will relate to a stable and tested version. There may be more recent builds
Tags follow these patterns:
:cuda-[x.x.x-base|runtime]-[ubuntu-version]:rocm-[x.x.x-runtime]-[ubuntu-version]:cpu-[ubuntu-version]Browse ghcr.io for an image suitable for your target environment. Alternatively, view a select range of CUDA and ROCm builds at DockerHub.
Supported Platforms: NVIDIA CUDA, AMD ROCm, CPU
| Variable | Description |
|---|---|
AUTO_UPDATE | Update ComfyUI on startup (default false) |
CIVITAI_TOKEN | Authenticate download requests from Civitai - Required for gated models |
COMFYUI_ARGS | Startup arguments. eg. --gpu-only --highvram |
COMFYUI_PORT_HOST | ComfyUI interface port (default 8188) |
COMFYUI_REF | Git reference for auto update. Accepts branch, tag or commit hash. Default: latest release |
COMFYUI_URL | Override $DIRECT_ADDRESS:port with URL for ComfyUI |
HF_TOKEN | Authenticate download requests from HuggingFace - Required for gated models (SD3, FLUX, etc.) |
See the base environment variables here for more configuration options.
| Environment | Packages |
|---|---|
comfyui | ComfyUI and dependencies |
api | ComfyUI API wrapper and dependencies |
The comfyui environment will be activated on shell login.
See the base micromamba environments here.
The following services will be launched alongside the default services provided by the base image.
The service will launch on port 8188 unless you have specified an override with COMFYUI_PORT_HOST.
You can set startup flags by using variable COMFYUI_ARGS.
To manage this service you can use supervisorctl [start|stop|restart] comfyui.
This service is available on port 8188 and is a work-in-progress to replace previous serverless handlers which have been depreciated; Old Docker images and sources remain available should you need them.
You can access the api directly at /ai-dock/api/ or you can use the Swager/openAPI playground at /ai-dock/api/docs.
[!NOTE] All services are password protected by default. See the security and environment variables documentation for more information.
Vast.ai
The author (@robballantyne) may be compensated if you sign up to services linked in this document. Testing multiple variants of GPU images in many different environments is both costly and time-consuming; This helps to offset costs
603 followers · starred Nov 2023
88 followers · starred Apr 2024
509 followers · starred Oct 2024
469 followers · starred Sep 2025
ComfyUI docker images for use in GPU cloud and local environments. Includes AI-Dock base for authentication and improved user experience.
Shell
1,067
111 commits
updated Nov 4, 2024
Run ComfyUI in a highly-configurable, cloud-first AI-Dock container.
[!NOTE] These images do not bundle models or third-party configurations. You should use a provisioning script to automatically configure your container. You can find examples, including
SD3&FLUX.1setup, inconfig/provisioning.
All AI-Dock containers share a common base which is designed to make running on cloud services such as vast.ai as straightforward and user friendly as possible.
Common features and options are documented in the base wiki but any additional features unique to this image will be detailed below.
The :latest tag points to :latest-cuda and will relate to a stable and tested version. There may be more recent builds
Tags follow these patterns:
:cuda-[x.x.x-base|runtime]-[ubuntu-version]:rocm-[x.x.x-runtime]-[ubuntu-version]:cpu-[ubuntu-version]Browse ghcr.io for an image suitable for your target environment. Alternatively, view a select range of CUDA and ROCm builds at DockerHub.
Supported Platforms: NVIDIA CUDA, AMD ROCm, CPU
| Variable | Description |
|---|---|
AUTO_UPDATE | Update ComfyUI on startup (default false) |
CIVITAI_TOKEN | Authenticate download requests from Civitai - Required for gated models |
COMFYUI_ARGS | Startup arguments. eg. --gpu-only --highvram |
COMFYUI_PORT_HOST | ComfyUI interface port (default 8188) |
COMFYUI_REF | Git reference for auto update. Accepts branch, tag or commit hash. Default: latest release |
COMFYUI_URL | Override $DIRECT_ADDRESS:port with URL for ComfyUI |
HF_TOKEN | Authenticate download requests from HuggingFace - Required for gated models (SD3, FLUX, etc.) |
See the base environment variables here for more configuration options.
| Environment | Packages |
|---|---|
comfyui | ComfyUI and dependencies |
api | ComfyUI API wrapper and dependencies |
The comfyui environment will be activated on shell login.
See the base micromamba environments here.
The following services will be launched alongside the default services provided by the base image.
The service will launch on port 8188 unless you have specified an override with COMFYUI_PORT_HOST.
You can set startup flags by using variable COMFYUI_ARGS.
To manage this service you can use supervisorctl [start|stop|restart] comfyui.
This service is available on port 8188 and is a work-in-progress to replace previous serverless handlers which have been depreciated; Old Docker images and sources remain available should you need them.
You can access the api directly at /ai-dock/api/ or you can use the Swager/openAPI playground at /ai-dock/api/docs.
[!NOTE] All services are password protected by default. See the security and environment variables documentation for more information.
Vast.ai
The author (@robballantyne) may be compensated if you sign up to services linked in this document. Testing multiple variants of GPU images in many different environments is both costly and time-consuming; This helps to offset costs
603 followers · starred Nov 2023
88 followers · starred Apr 2024
509 followers · starred Oct 2024
469 followers · starred Sep 2025