∇ Prompt

This is the biggest node in the whole pack, field-count-wise, and it's built around a real fact about Illustrious: it reads Danbooru's tag vocabulary as its native interface, and that vocabulary covers a lot more than subject matter - camera angle, framing, lighting, focus, all as standardized tags the model actually learned during training, not free description. This node turns that vocabulary into dropdowns instead of something you have to already know.

What's actually in here

Shot Type, Framing, Perspective, Focus, Shot Angle, Lighting, Depth & Focus, and Motion & Expression are all composition and camera controls, and their tooltips are specific about it - Focus, for instance, is labeled "Focus terms trained into Illustrious models," meaning these aren't invented categories, they're real tags the model has seen enough of to respond to reliably (sharp focus, shallow dof, from above, dutch angle, and so on). Girl Characters/Boy Characters inject the counting convention Danbooru-lineage prompts use - 1girl solo, 2 girls - rather than making you type it, and Mature Characters swaps that wording to woman/man.

Character and Artist (plus their E621 counterparts) are free-text fields that hit the pack's own live search - the same 5,000+ name character/artist database the README describes, built specifically because artist tags are one of the things Illustrious does that a lot of other models can't: they're a real style lever here, not decoration.

Negative Prompt is itself a dropdown of curated presets (Basic / Extra / Boost / Article (Recommended) / Latest (Underscores)) rather than a text field - a shortcut past typing out a negative tag list by hand every time, and generally a reasonable one, since a short focused negative tends to outperform a forty-keyword wall on this model family anyway.

Quality Boost (Beta), Quality Style, Quality Variation, Prefix QB (Beta), and Cinematic (Beta) are all marked Beta in their own tooltips - the pack is explicitly still in beta overall (its author has been open on Reddit that interfaces and defaults shift release to release), so treat these five as more likely to change than the rest of the node.

Format Tag wraps output in escaped parentheses for weight emphasis, matching Danbooru/Illustrious-style syntax; Break Format splits the prompt into segments for UIs that reset token budget at a break. Add Chain Insert places a placeholder token for a downstream chained prompt node to fill in - this is how you compose this node with Characters or Artists into one final string without a separate string-concatenation step.

The inputs and outputs that matter

Nearly every field has a sensible - (off) default, so you only need to touch the ones you actually want. Prompt itself is the one free-text field, pre-filled with an example so the node isn't blank on first drop.

Outputs: PROMPT and NEGATIVE, both strings - wire straight into your positive and negative CLIP Text Encode nodes.

Installing it

ComfyUI Manager: search "EasyIllustrious," install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/regiellis/ComfyUI-EasyIllustrious
pip install -r ComfyUI-EasyIllustrious/requirements.txt

No model downloads - the character/artist search runs against the pack's own bundled database, not an external service.

Common issues

Character or artist search comes back empty. Match the README's own troubleshooting: keep your query at least 2 characters, and use underscore format (hatsune_miku, not "hatsune miku") - the field's own tooltip examples (saber, nezuko, kantoku, wlop) already follow that convention.

SFW toggle isn't actually keeping content clean. It's a prompt-level nudge, not a hard filter - pair it with your checkpoint choice and negative prompt rather than relying on it alone for anything that actually matters.