Fusion 360 Image Output
Actually write those generated images to the Fusion 360 asset path
- images
- written_paths_json
- primary_path
- summary_json
The runtime half of the Fusion 360 image round-trip. MKRFusion360ImageOutputPlan decides where an image should go and what it should be; MKRFusion360ImageOutput is the node that takes your generated IMAGE batch and actually writes the files to the planned asset path. Plan and execute, cleanly separated.
It belongs to the pack's shared host-image-output family, so its behavior is identical to the Blender, Nuke, Photoshop, After Effects, Premiere, and Maya image-output nodes - same base class, different host label. Once you've used one, you've used them all, which is a big part of the appeal of this lane.
How it works
Two required inputs:
images- the IMAGE tensor batch you generated. This is what gets written to disk.image_output_plan_json- the plan fromMKRFusion360ImageOutputPlan(themkrshift_fusion360_image_output_plan_v1schema). This carries the targetasset_path,apply_modeandimage_role. The node reads the path out of the plan and writes there.
One optional: filename_override - a string that overrides the planned filename if you need a different stem on the way out. Leave empty to follow the plan.
What comes back:
written_paths_json- JSON with thepathsarray (every file actually written) plus the plan it used. This is your receipt.primary_path- the first written path, as a plain string. The easy one to wire into a manifest or log.summary_json- host, apply mode, image role, file count, primary path, and warnings.
How it fits
The flow you'll actually build: generate → MKRFusion360ImageOutputPlan (decide) → MKRFusion360ImageOutput (write) → the Fusion 360 add-in picks the file up and applies it as the decal/appearance/canvas the plan described. Nothing here talks to Fusion directly - it writes files to the paths the add-in expects. If the add-in isn't picking up your image, the first things to check are the plan's asset_path and that the file landed where primary_path says it did.
Installing
Same story as every node in this pack:
cd ComfyUI/custom_nodes
git clone https://github.com/criskb/MKRShift_Nodes
Restart ComfyUI, or ComfyUI Manager search "MKRShift_Nodes". No extra Python deps. The Fusion 360 add-in from addons/fusion360/ in the repo is the other half of the loop.
The honest framing: this node is simple by design - parse the plan, write the images, report the paths. Its entire value is being the correct, predictable "write" step in a CAD round-trip pipeline, and in being identical in shape to the other host output nodes so your muscle memory transfers. It does nothing clever, and for a file-writing node, that's exactly what you want. There's no community chatter to cite yet - the pack is young - but the plan/execute split here is a pattern that scales past Fusion 360 into every other host bridge the pack ships.
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| images | IMAGE | — | |
| image_output_plan_json | STRING | {} | — |
| filename_overrideopt | STRING | — |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| written_paths_json | STRING | — |
| primary_path | STRING | — |
| summary_json | STRING | — |