Submit ComfyUI workflows to Thinkbox Deadline render farm.
Check out a quick demo showing how ComfyUI can play nice with Deadline. Please check out the video:
- Submit ComfyUI workflows directly to Deadline
- Batch rendering with seed variation
- Real-time progress monitoring via Deadline Monitor
- Configurable pools, groups, and priorities
- Open ComfyUI Manager → Install Custom Nodes
- Search "ComfyUI Deadline Submission" → Install
- Restart ComfyUI
cd ComfyUI/custom_nodes
git clone https://github.com/YOUR_USERNAME/ComfyUI-Deadline-Plugin.gitCopy plugins/ComfyUI/ to your Deadline Repository's custom/plugins/ directory and restart Deadline services.
- Add "Submit to Deadline" node to your workflow
- Configure job settings (name, priority, pool, etc.)
- Execute workflow
- Monitor progress in Deadline Monitor
- batch_count: Number of tasks (1-100)
- change_seeds_per_task: Randomize seeds for different outputs
- priority: Job priority (0-100)
- pool/group: Deadline worker assignment
For render farms with shared storage, copy example_extra_model_paths.yaml to your ComfyUI installation as extra_model_paths.yaml and update paths.
In Deadline Monitor → Repository Options → ComfyUI plugin configuration, the ComfyUI Installation Paths field now accepts multiple entries. Add one path per line (drive letters, UNC shares, etc.). Workers will walk the list from top to bottom, apply Deadline path mapping, expand environment variables, and pick the first path that contains both ComfyUI/main.py and python_embeded/python.exe. This mirrors the multi-path behavior used by Deadline's Houdini plugin.
Example:
C:\ComfyUI_windows_portable
D:\Apps\ComfyUI
\\NAS01\software\ComfyUI
- Captures current ComfyUI workflow
- Submits to Deadline with proper configuration
- Workers execute workflow via ComfyUI API
- Progress reported through Deadline Monitor
- ComfyUI installation on worker machines
- Thinkbox Deadline
- No additional Python dependencies (uses standard library)