Set Dict Conditioning

The dict nodes are antrobot's answer to wire spaghetti, and they take a different route than pipes. Instead of bundling a fixed set of things (like a BASIC_PIPE always being model/CLIP/VAE/pos/neg), a DICT is an open bag: you stash any value under a name you choose, carry the bag around on one wire, and pull things back out by name later. Set Dict Conditioning is the version that stashes a CONDITIONING - an encoded prompt - into that bag.

Why bother? Say you've got three different prompt conditionings you want available deep in a workflow - a base prompt, a style prompt, a detail prompt. Rather than running three wires across the whole graph, you setDict each one under a key ("base", "style", "detail"), pass the single DICT along, and getDict whichever you need where you need it. It's the same instinct as rgthree's Context nodes - carry a bundle on one wire - but with names you pick instead of fixed slots.

How it works

Per the README, Set Dict is the core of the whole system: any value you put in gets packaged into a DICT labeled with your key and sent out. Crucially, you can also pass a DICT in - and then it appends your new value to that existing dict instead of starting a fresh one. That's how you build up a bag with several entries: chain Set Dict nodes, each adding one more key, passing the growing DICT down the line.

The inputs and outputs that matter

• key (required, text) - the name you'll retrieve this conditioning by later. Keep it memorable and consistent; you type the exact same string into the matching Get Dict node.
• CONDITIONING (required) - the encoded prompt to store.
• DICT (optional) - an existing dict to add to. Leave it empty to start a new one; wire in the previous Set Dict's output to keep appending.

One output: DICT, the bag with your entry added. Chain it into the next Set Dict, or send it wherever the conditioning needs to be unpacked.

Installing it

Pure Python, no model downloads, nothing heavy - quick install. ComfyUI Manager → Install Custom Nodes → search "antrobots ComfyUI Nodepack" → Install → restart, or clone it:

cd ComfyUI/custom_nodes
git clone https://github.com/antrobot1234/antrobots-comfyUI-nodepack

then restart. It's under antrobots-ComfyUI-nodepack/dicts/set.

Common issues

The one thing that will bite you: keys are just strings, and typos are silent. If you setDict under "positive" and later getDict for "postive", the get won't find it and quietly falls back to its default - no error, just the wrong conditioning. When a dict-based workflow does something inexplicable, check your key spelling on both ends first.

Also remember to actually chain the DICT input when you want multiple entries. Forget to wire the previous dict in and each Set Dict starts fresh, so you end up with a bag holding only the last thing you set. And retrieval is by the matching Get Dict Conditioning node, not a stock node - the DICT type only talks to these dict nodes. It's a small solo pack; real bugs go to the GitHub issue tracker, which the author explicitly welcomes.