Camera Image Pipe Loader
The camera pipe loader
- model
- positive
- negative
- latent
- vae
- clip
- pipe
Here's the ComfyUI reality everyone hits eventually: your workflow grows past one generation and suddenly the KSampler has seven inputs and you're dragging wires across the whole canvas to reach them. The fix this pack uses is a pipe - bundle model, positive, negative, latent, vae, clip, filename, and seed into one CAMERA_PIPE_LINE wire, and push a single connection down the chain instead of eight. CameraImagePipeLoader is the node that creates that bundle.
It does nothing clever. The description is upfront: "This node does not perform any processing on the inputs, it simply passes them through as a single object." You connect what you have, it packs it, and out comes one pipe. The point isn't the node - it's the topology it enables.
How it works
Every input is optional, because a pipe can be built incrementally. Feed it a model and a latent, get a pipe holding just those; the empty slots stay empty until you fill them with CameraImagePipeComposer downstream. The inputs are the full sampling state the pack expects in a camera pipe:
model(MODEL),positive/negative(CONDITIONING),latent(LATENT),vae(VAE),clip(CLIP)filename(STRING),seed(INT, full 64-bit range)
The output is a single pipe of type CAMERA_PIPE_LINE. From there the pack's pipe-compatible nodes can carry it along - and at the end, CameraImagePipeDecomposer unpacks it into the individual wires you feed to a real sampler.
When it earns its keep
If you're building one image, skip it - you're just adding a hop. It pays off when you have a reusable "camera roll" subgraph or a batch/experiment setup where you swap one component (the seed, say) while keeping the rest. Pipe the bundle through, swap a single field with the Composer, and re-run without re-plumbing. It's the same organizational instinct as the pipe utilities in packs like rgthree-comfy, just with the film-pipeline's own tuple layout.
Install
It's part of the pack - one install covers every node here:
cd ComfyUI/custom_nodes
git clone https://github.com/AlbertJBurton/comfyui-jbnodes.git
cd comfyui-jbnodes
pip install -r requirements.txt
Restart ComfyUI, or search "comfyui-jbnodes" in ComfyUI Manager. Dependencies: numpy, torch, colour-science, scipy, Pillow.
Gotchas
- The pipe tuple is specific to this pack. Other packs' pipe nodes won't understand a
CAMERA_PIPE_LINE, and mixing them up is the #1 way people get "cannot connect" errors with these utilities. - Because the pack is beta, the pipe's internal layout can change between versions - a workflow that loads a raw workflow JSON from before an update may fail to wire until you reconnect the Loader.
- No model files, no downloads. This node is pure tuple bookkeeping; the heavy lifting is elsewhere in the pack.
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| modelopt | MODEL | — | |
| positiveopt | CONDITIONING | — | |
| negativeopt | CONDITIONING | — | |
| latentopt | LATENT | — | |
| vaeopt | VAE | — | |
| clipopt | CLIP | — | |
| filenameopt | STRING | — | |
| seedopt | INT | 00–18446744073709550000 | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| pipe | CAMERA_PIPE_LINE | — |